Datasets:

sswt

/

arxiv_sample_50K

like 0

iVPF Numerical Invertible Volume Preserving Flow for Efficient Lossless Compression ; It is nontrivial to store rapidly growing big data nowadays, which demands highperformance lossless compression techniques. Likelihoodbased generative models have witnessed their success on lossless compression, where flow based models are desirable in allowing exact data likelihood optimisation with bijective mappings. However, common continuous flows are in contradiction with the discreteness of coding schemes, which requires either 1 imposing strict constraints on flow models that degrades the performance or 2 coding numerous bijective mapping errors which reduces the efficiency. In this paper, we investigate volume preserving flows for lossless compression and show that a bijective mapping without error is possible. We propose Numerical Invertible Volume Preserving Flow iVPF which is derived from the general volume preserving flows. By introducing novel computation algorithms on flow models, an exact bijective mapping is achieved without any numerical error. We also propose a lossless compression algorithm based on iVPF. Experiments on various datasets show that the algorithm based on iVPF achieves stateoftheart compression ratio over lightweight compression algorithms.

Dynamically polarisable forcefields for surface simulations via multioutput classification Neural Networks ; We present a general procedure to introduce electronic polarization into classical Molecular Dynamics MD forcefields using a Neural Network NN model. We apply this framework to the simulation of a solidliquid interface where the polarization of the surface is essential to correctly capture the main features of the system. By introducing a multiinput, multioutput NN and treating the surface polarization as a discrete classification problem, for which NNs are known to excel, we are able to obtain very good accuracy in terms of quality of predictions. Through the definition of a custom loss function we are able to impose a physically motivated constraint within the NN itself making this model extremely versatile, especially in the modelling of different surface charge states. The NN is validated considering the redistribution of electronic charge density within a graphene based electrode in contact with aqueous electrolyte solution, a system highly relevant to the development of next generation lowcost supercapacitors. We compare the performances of our NNMD model against Quantum MechanicsMolecular dynamics simulations where we obtain a most satisfactorily agreement.

Online Learning of a Probabilistic and Adaptive Scene Representation ; Constructing and maintaining a consistent scene model onthefly is the core task for online spatial perception, interpretation, and action. In this paper, we represent the scene with a Bayesian nonparametric mixture model, seamlessly describing perpoint occupancy status with a continuous probability density function. Instead of following the conventional data fusion paradigm, we address the problem of online learning the process how sequential point cloud data are generated from the scene geometry. An incremental and parallel inference is performed to update the parameter space in realtime. We experimentally show that the proposed representation achieves stateoftheart accuracy with promising efficiency. The consistent probabilistic formulation assures a generative model that is adaptive to different sensor characteristics, and the model complexity can be dynamically adjusted onthefly according to different data scales.

Analysis on Image Set Visual Question Answering ; We tackle the challenge of Visual Question Answering in multiimage setting for the ISVQA dataset. Traditional VQA tasks have focused on a singleimage setting where the target answer is generated from a single image. Image set VQA, however, comprises of a set of images and requires finding connection between images, relate the objects across images based on these connections and generate a unified answer. In this report, we work with 4 approaches in a bid to improve the performance on the task. We analyse and compare our results with three baseline models LXMERT, HMEVideoQA and VisualBERT and show that our approaches can provide a slight improvement over the baselines. In specific, we try to improve on the spatial awareness of the model and help the model identify color using enhanced pretraining, reduce language dependence using adversarial regularization, and improve counting using regression loss and graph based deduplication. We further delve into an indepth analysis on the language bias in the ISVQA dataset and show how models trained on ISVQA implicitly learn to associate language more strongly with the final answer.

Storchastic A Framework for General Stochastic Automatic Differentiation ; Modelers use automatic differentiation AD of computation graphs to implement complex Deep Learning models without defining gradient computations. Stochastic AD extends AD to stochastic computation graphs with sampling steps, which arise when modelers handle the intractable expectations common in Reinforcement Learning and Variational Inference. However, current methods for stochastic AD are limited They are either only applicable to continuous random variables and differentiable functions, or can only use simple but high variance scorefunction estimators. To overcome these limitations, we introduce Storchastic, a new framework for AD of stochastic computation graphs. Storchastic allows the modeler to choose from a wide variety of gradient estimation methods at each sampling step, to optimally reduce the variance of the gradient estimates. Furthermore, Storchastic is provably unbiased for estimation of anyorder gradients, and generalizes variance reduction techniques to higherorder gradient estimates. Finally, we implement Storchastic as a PyTorch library at httpsgithub.comHEmilestorchastic.

Deepfake Detection Scheme Based on Vision Transformer and Distillation ; Deepfake is the manipulated video made with a generative deep learning technique such as Generative Adversarial Networks GANs or Auto Encoder that anyone can utilize. Recently, with the increase of Deepfake videos, some classifiers consisting of the convolutional neural network that can distinguish fake videos as well as deepfake datasets have been actively created. However, the previous studies based on the CNN structure have the problem of not only overfitting, but also considerable misjudging fake video as real ones. In this paper, we propose a Vision Transformer model with distillation methodology for detecting fake videos. We design that a CNN features and patchbased positioning model learns to interact with all positions to find the artifact region for solving false negative problem. Through comparative analysis on Deepfake Detection DFDC Dataset, we verify that the proposed scheme with patch embedding as input outperforms the stateoftheart using the combined CNN features. Without ensemble technique, our model obtains 0.978 of AUC and 91.9 of f1 score, while previous SOTA model yields 0.972 of AUC and 90.6 of f1 score on the same condition.

MMBERT Multimodal BERT Pretraining for Improved Medical VQA ; Images in the medical domain are fundamentally different from the general domain images. Consequently, it is infeasible to directly employ general domain Visual Question Answering VQA models for the medical domain. Additionally, medical images annotation is a costly and timeconsuming process. To overcome these limitations, we propose a solution inspired by selfsupervised pretraining of Transformerstyle architectures for NLP, Vision and Language tasks. Our method involves learning richer medical image and text semantic representations using Masked Language Modeling MLM with image features as the pretext task on a large medical imagecaption dataset. The proposed solution achieves new stateoftheart performance on two VQA datasets for radiology images VQAMed 2019 and VQARAD, outperforming even the ensemble models of previous best solutions. Moreover, our solution provides attention maps which help in model interpretability. The code is available at httpsgithub.comVirajBagalMMBERT

Adaptive rewiring of random neural networks generates convergentdivergent units ; Brain networks are adaptively rewired continually, adjusting their topology to bring about functionality and efficiency in sensory, motor and cognitive tasks. In model neural network architectures, adaptive rewiring generates complex, brainlike topologies. Present models, however, cannot account for the emergence of complex directed connectivity structures. We tested a biologically plausible model of adaptive rewiring in directed networks, based on two algorithms widely used in distributed computing advection and consensus. When both are used in combination as rewiring criteria, adaptive rewiring shortens path length and enhances connectivity. When keeping a balance between advection and consensus, adaptive rewiring produces convergentdivergent units consisting of convergent hub nodes, which collect inputs from pools of sparsely connected, or local, nodes and project them via densely interconnected processing nodes onto divergent hubs that broadcast output back to the local pools. Convergentdivergent units operate within and between sensory, motor, and cognitive brain regions as their connective core, mediating contextsensitivity to local network units. By showing how these structures emerge spontaneously in directed networks models, adaptive rewiring offers selforganization as a principle for efficient information propagation and integration in the brain.

SemiSupervised Clustering with Inaccurate Pairwise Annotations ; Pairwise relational information is a useful way of providing partial supervision in domains where class labels are difficult to acquire. This work presents a clustering model that incorporates pairwise annotations in the form of mustlink and cannotlink relations and considers possible annotation inaccuracies i.e., a common setting when experts provide pairwise supervision. We propose a generative model that assumes Gaussiandistributed data samples along with mustlink and cannotlink relations generated by stochastic block models. We adopt a maximumlikelihood approach and demonstrate that, even when supervision is weak and inaccurate, accounting for relational information significantly improves clustering performance. Relational information also helps to detect meaningful groups in realworld datasets that do not fit the original datadistribution assumptions. Additionally, we extend the model to integrate prior knowledge of experts' accuracy and discuss circumstances in which the use of this knowledge is beneficial.

Nutribullets Hybrid Multidocument Health Summarization ; We present a method for generating comparative summaries that highlights similarities and contradictions in input documents. The key challenge in creating such summaries is the lack of large parallel training data required for training typical summarization systems. To this end, we introduce a hybrid generation approach inspired by traditional concepttotext systems. To enable accurate comparison between different sources, the model first learns to extract pertinent relations from input documents. The content planning component uses deterministic operators to aggregate these relations after identifying a subset for inclusion into a summary. The surface realization component lexicalizes this information using a textinfilling language model. By separately modeling content selection and realization, we can effectively train them with limited annotations. We implemented and tested the model in the domain of nutrition and health rife with inconsistencies. Compared to conventional methods, our framework leads to more faithful, relevant and aggregationsensitive summarization while being equally fluent.

Neural Temporal Point Processes A Review ; Temporal point processes TPP are probabilistic generative models for continuoustime event sequences. Neural TPPs combine the fundamental ideas from point process literature with deep learning approaches, thus enabling construction of flexible and efficient models. The topic of neural TPPs has attracted significant attention in the recent years, leading to the development of numerous new architectures and applications for this class of models. In this review paper we aim to consolidate the existing body of knowledge on neural TPPs. Specifically, we focus on important design choices and general principles for defining neural TPP models. Next, we provide an overview of application areas commonly considered in the literature. We conclude this survey with the list of open challenges and important directions for future work in the field of neural TPPs.

Permutation invariant Gaussian 2matrix models ; We construct the general permutation invariant Gaussian 2matrix model for matrices of arbitrary size D. The parameters of the model are given in terms of variables defined using the representation theory of the symmetric group SD. A correspondence is established between the permutation invariant polynomial functions of the matrix variables the observables of the model and directed colored graphs, which sheds light on stability properties in the large D counting of these invariants. The refined counting of the graphs is given in terms of double cosets involving permutation groups defined by the local structure of the graphs. Linear and quadratic observables are transformed to an SD representation theoretic basis and are used to define the convergent Gaussian measure. The perturbative rules for the computation of expectation values of graphbasis observables of any degree are given in terms of the representation theoretic parameters. Explicit results for a number of observables of degree up to four are given along with a Sage programme that computes general expectation values.

Congruence and Plausibility, not Presence Pivotal Conditions for XR Experiences and Effects, a Novel Model ; Presence often is considered the most important quale describing the subjective feeling of being in a computergenerated andor computermediated virtual environment. The identification and separation of orthogonal presence components, i.e., the place illusion and the plausibility illusion, has been an accepted theoretical model describing Virtual Reality VR experiences for some time. This perspective article challenges this presenceoriented VR theory. First, we argue that a place illusion cannot be the major construct to describe the much wider scope of Virtual, Augmented, and Mixed Reality VR, AR, MR or XR for short. Second, we argue that there is no plausibility illusion but merely plausibility, and we derive the place illusion caused by congruent and plausible generation of spatial cues, and similarly for all the current model's sodefined illusions. Finally, we propose congruence and plausibility to become the central essential conditions in a novel theoretical model describing XR experiences and effects.

On the Inductive Bias of Masked Language Modeling From Statistical to Syntactic Dependencies ; We study how masking and predicting tokens in an unsupervised fashion can give rise to linguistic structures and downstream performance gains. Recent theories have suggested that pretrained language models acquire useful inductive biases through masks that implicitly act as cloze reductions for downstream tasks. While appealing, we show that the success of the random masking strategy used in practice cannot be explained by such clozelike masks alone. We construct clozelike masks using taskspecific lexicons for three different classification datasets and show that the majority of pretrained performance gains come from generic masks that are not associated with the lexicon. To explain the empirical success of these generic masks, we demonstrate a correspondence between the Masked Language Model MLM objective and existing methods for learning statistical dependencies in graphical models. Using this, we derive a method for extracting these learned statistical dependencies in MLMs and show that these dependencies encode useful inductive biases in the form of syntactic structures. In an unsupervised parsing evaluation, simply forming a minimum spanning tree on the implied statistical dependence structure outperforms a classic method for unsupervised parsing 58.74 vs. 55.91 UUAS.

DocumentLevel Event Argument Extraction by Conditional Generation ; Event extraction has long been treated as a sentencelevel task in the IE community. We argue that this setting does not match human informationseeking behavior and leads to incomplete and uninformative extraction results. We propose a documentlevel neural event argument extraction model by formulating the task as conditional generation following event templates. We also compile a new documentlevel event extraction benchmark dataset WikiEvents which includes complete event and coreference annotation. On the task of argument extraction, we achieve an absolute gain of 7.6 F1 and 5.7 F1 over the next best model on the RAMS and WikiEvents datasets respectively. On the more challenging task of informative argument extraction, which requires implicit coreference reasoning, we achieve a 9.3 F1 gain over the best baseline. To demonstrate the portability of our model, we also create the first endtoend zeroshot event extraction framework and achieve 97 of fully supervised model's trigger extraction performance and 82 of the argument extraction performance given only access to 10 out of the 33 types on ACE.

Constrained Language Models Yield FewShot Semantic Parsers ; We explore the use of large pretrained language models as fewshot semantic parsers. The goal in semantic parsing is to generate a structured meaning representation given a natural language input. However, language models are trained to generate natural language. To bridge the gap, we use language models to paraphrase inputs into a controlled sublanguage resembling English that can be automatically mapped to a target meaning representation. Our results demonstrate that with only a small amount of data and very little code to convert into Englishlike representations, our blueprint for rapidly bootstrapping semantic parsers leads to surprisingly effective performance on multiple community tasks, greatly exceeding baseline methods also trained on the same limited data.

Sensitivity Analysis of Passenger Behavioral Model for Dynamic Pricing of Shared Mobility on Demand ; This paper provides a framework to quantify the sensitivity associated with behavioral models based on Cumulative Prospect Theory CPT. These are used to design dynamic pricing strategies aimed at maximizing performance metrics of the Shared Mobility OnDemand Service SMoDS, as solutions to a constrained nonlinear optimization problem. We analyze the sensitivity of both the optimal tariff as well as the optimal objective function with respect to CPT model parameters. In addition to deriving analytical solutions under certain assumptions, more general numerical results are obtained via computational experiments and simulations to analyze the sensitivity. We find that the model is relatively robust for small to moderate parameter perturbations. Although some of the trends in sensitivity are fairly general, the exact nature of variations in many cases depends heavily on the specific travel scenarios and modes being considered. This is primarily due to the complex nonlinearities in the problem, as well as the significant heterogeneity in passenger preferences across different types of trips.

Federated Traffic Synthesizing and Classification Using Generative Adversarial Networks ; With the fast growing demand on new services and applications as well as the increasing awareness of data protection, traditional centralized traffic classification approaches are facing unprecedented challenges. This paper introduces a novel framework, Federated Generative Adversarial Networks and Automatic Classification FGANAC, which integrates decentralized data synthesizing with traffic classification. FGANAC is able to synthesize and classify multiple types of service data traffic from decentralized local datasets without requiring a large volume of manually labeled dataset or causing any data leakage. Two types of data synthesizing approaches have been proposed and compared computationefficient FGAN FGANuppercaseexpandafterromannumeral1 and communicationefficient FGAN FGANuppercaseexpandafterromannumeral2. The former only implements a single CNN model for processing each local dataset and the later only requires coordination of intermediate model training parameters. An automatic data classification and model updating framework has been proposed to automatically identify unknown traffic from the synthesized data samples and create new pseudolabels for model training. Numerical results show that our proposed framework has the ability to synthesize highly mixed service data traffic and can significantly improve the traffic classification performance compared to existing solutions.

KXLNet A General Method for Combining Explicit Knowledge with Language Model Pretraining ; Though pretrained language models such as Bert and XLNet, have rapidly advanced the stateoftheart on many NLP tasks, they implicit semantics only relying on surface information between words in corpus. Intuitively, background knowledge influences the efficacy of understanding. Inspired by this common sense, we focus on improving model pretraining by leveraging explicit knowledge. Different from recent research that optimize pretraining model by knowledge masking strategies, we propose a simple but general method to combine explicit knowledge with pretraining. To be specific, we first match knowledge facts from knowledge graph KG and then add a knowledge injunction layer to transformer directly without changing its architecture. The present study seeks to find the direct impact of explicit knowledge on transformer pertraining. We conduct experiments on various datasets for different downstream tasks. The experimental results show that solely by adding external knowledge to transformer can improve the learning performance on many NLP tasks.

Quillen equivalences inducing Grothendieck duality for unbounded chain complexes of sheaves ; Let mathbbX be a semiseparated Noetherian scheme with a dualizing complex D. We lift some wellknown triangulated equivalences associated with Grothendieck duality to Quillen equivalences of model categories. In the process we are able to show that the Gorenstein flat model structure, on the category of quasicoherent sheaves on mathbbX, is Quillen equivalent to the Gorenstein injective model structure. Also noteworthy is that we extend the recollement of Krause to hold without the Noetherian condition. Using a set of flat generators, it holds for any quasicompact semiseparated scheme mathbbX. With this we also show that the Gorenstein injective quasicoherent sheaves are the fibrant objects of a cofibrantly generated abelian model structure for any semiseparated Noetherian scheme mathbbX. Finally, we consider both the injective and mock projective approach to Tate cohomology of quasicoherent sheaves. They agree whenever mathbbX is a semiseparated Gorenstein scheme of finite Krull dimension.

A Robustness Analysis of Inverse Optimal Control of Bipedal Walking ; Cost functions have the potential to provide compact and understandable generalizations of motion. The goal of Inverse Optimal Control IOC is to analyze an observed behavior which is assumed to be optimal with respect to an unknown cost function, and infer this cost function. Here we develop a method for characterizing cost functions of legged locomotion, with the goal of representing complex humanoid behavior with simple models. To test this methodology we simulate walking gaits of a simple 5 link planar walking model which optimize known cost functions, and assess the ability of our IOC method to recover them. In particular, the IOC method uses an iterative trajectory optimization process to infer cost function weightings consistent with those used to generate a single demonstrated optimal trial. We also explore sensitivity of the IOC to sensor noise in the observed trajectory, imperfect knowledge of the model or task, as well as uncertainty in the components of the cost function used. With appropriate modeling, these methods may help infer cost functions from human data, yielding a compact and generalizable representation of humanlike motion for use in humanoid robot controllers, as well as providing a new tool for experimentally exploring human preferences.

Diverse Image Inpainting with Bidirectional and Autoregressive Transformers ; Image inpainting is an underdetermined inverse problem, which naturally allows diverse contents to fill up the missing or corrupted regions realistically. Prevalent approaches using convolutional neural networks CNNs can synthesize visually pleasant contents, but CNNs suffer from limited perception fields for capturing global features. With imagelevel attention, transformers enable to model longrange dependencies and generate diverse contents with autoregressive modeling of pixelsequence distributions. However, the unidirectional attention in autoregressive transformers is suboptimal as corrupted image regions may have arbitrary shapes with contexts from any direction. We propose BATFill, an innovative image inpainting framework that introduces a novel bidirectional autoregressive transformer BAT for image inpainting. BAT utilizes the transformers to learn autoregressive distributions, which naturally allows the diverse generation of missing contents. In addition, it incorporates the masked language model like BERT, which enables bidirectionally modeling of contextual information of missing regions for better image completion. Extensive experiments over multiple datasets show that BATFill achieves superior diversity and fidelity in image inpainting qualitatively and quantitatively.

Dynamic VAEs with Generative Replay for Continual Zeroshot Learning ; Continual zeroshot learningCZSL is a new domain to classify objects sequentially the model has not seen during training. It is more suitable than zeroshot and continual learning approaches in realcase scenarios when data may come continually with only attributes for a few classes and attributes and features for other classes. Continual learningCL suffers from catastrophic forgetting, and zeroshot learningZSL models cannot classify objects like stateoftheart supervised classifiers due to lack of actual dataor features during training. This paper proposes a novel continual zeroshot learning DVGRCZSL model that grows in size with each task and uses generative replay to update itself with previously learned classes to avoid forgetting. We demonstrate our hybrid modelDVGRCZSL outperforms the baselines and is effective on several datasets, i.e., CUB, AWA1, AWA2, and aPY. We show our method is superior in task sequentially learning with ZSLZeroShot Learning. We also discuss our results on the SUN dataset.

Nonsingular cosmological models with strong gravity in the past ; In scalartensor Horndeski theories, nonsingular cosmological models bounce and genesis are problematic because of potential ghost andor gradient instabilities. One way to get around this obstacle is to send the effective Planck mass to zero in the asymptotic past strong gravity in the past. One may suspect that this feature is a signal of a strong coupling problem at early times. However, the classical treatment of the cosmological background is legitimate, provided that the strong coupling energy scale remains at all times much higher than the scale associated with the classical evolution. We construct various models of this sort, namely i bouncing Universe which proceeds through inflationary epoch to kination expansion within general relativity, driven by massless scalar field; ii bouncing Universe with kination stage immediately after bounce; iii combination of genesis and bounce, with the Universe starting from flat spacetime, then contracting and bouncing to the expansion epoch; iv standard genesis evading the strong coupling problem in the past. All these models are stable, and perturbations about the backgrounds are not superluminal.

Improving BERT Model Using Contrastive Learning for Biomedical Relation Extraction ; Contrastive learning has been used to learn a highquality representation of the image in computer vision. However, contrastive learning is not widely utilized in natural language processing due to the lack of a general method of data augmentation for text data. In this work, we explore the method of employing contrastive learning to improve the text representation from the BERT model for relation extraction. The key knob of our framework is a unique contrastive pretraining step tailored for the relation extraction tasks by seamlessly integrating linguistic knowledge into the data augmentation. Furthermore, we investigate how largescale data constructed from the external knowledge bases can enhance the generality of contrastive pretraining of BERT. The experimental results on three relation extraction benchmark datasets demonstrate that our method can improve the BERT model representation and achieve stateoftheart performance. In addition, we explore the interpretability of models by showing that BERT with contrastive pretraining relies more on rationales for prediction. Our code and data are publicly available at httpsgithub.comudelbiotmlabBERTCLRE.

First evidence that nonmetricity fQ gravity could challenge CDM ; We propose a novel model in the framework of fQ gravity, which is a gravitational modification class arising from the incorporation of nonmetricity. The model has General Relativity as a particular limit, it has the same number of free parameters to those of LambdaCDM, however at a cosmological framework it gives rise to a scenario that does not have LambdaCDM as a limit. Nevertheless, confrontation with observations at both background and perturbation levels, namely with Supernovae type Ia SNIa, Baryonic Acoustic Oscillations BAO, cosmic chronometers CC, and Redshift Space Distortion RSD data, reveals that the scenario, according to AIC, BIC and DIC information criteria, is in some datasets slightly preferred comparing to LambdaCDM cosmology, although in all cases the two models are statistically indiscriminate. Finally, the model does not exhibit early dark energy features, and thus it immediately passes BBN constraints, while the variation of the effective Newton's constant lies well inside the observational bounds.

Nonlinear optical processes in cavity lightmatter systems ; We study nonlinear optical effects in electron systems with and without inversion symmetry in a FabryPerot cavity. General photon up and downconversion processes are modeled by the coupling of a noninteracting lattice model to two modes of the quantized light field. Effective descriptions retaining the most relevant states are devised via downfolding and a generalized Householder transformation. These models are used to relate the transition amplitudes for even order photonconversion processes to the shift vector, a topological quantity describing the difference in polarization between the valence and conduction band in noncentrosymmetric systems. We also demonstrate that the truncated models, despite their small Hilbert space, capture correlation effects induced by the photons in the electronic subsystem.

Estimating the conditional distribution in functional regression problems ; We consider the problem of consistently estimating the conditional distribution PY in A X of a functional data object YYt tin0,1 given covariates X in a general space, assuming that Y and X are related by a functional linear regression model. Two natural estimation methods are proposed, based on either bootstrapping the estimated model residuals, or fitting functional parametric models to the model residuals and estimating PY in A X via simulation. Whether either of these methods lead to consistent estimation depends on the consistency properties of the regression operator estimator, and the space within which Y is viewed. We show that under general consistency conditions on the regression operator estimator, which hold for certain functional principal component based estimators, consistent estimation of the conditional distribution can be achieved, both when Y is an element of a separable Hilbert space, and when Y is an element of the Banach space of continuous functions. The latter results imply that sets A that specify path properties of Y, which are of interest in applications, can be considered. The proposed methods are studied in several simulation experiments, and data analyses of electricity price and pollution curves.

A First Look Towards Explainable TextVQA Models via Visual and Textual Explanations ; Explainable deep learning models are advantageous in many situations. Prior work mostly provide unimodal explanations through posthoc approaches not part of the original system design. Explanation mechanisms also ignore useful textual information present in images. In this paper, we propose MTXNet, an endtoend trainable multimodal architecture to generate multimodal explanations, which focuses on the text in the image. We curate a novel dataset TextVQAX, containing ground truth visual and multireference textual explanations that can be leveraged during both training and evaluation. We then quantitatively show that training with multimodal explanations complements model performance and surpasses unimodal baselines by up to 7 in CIDEr scores and 2 in IoU. More importantly, we demonstrate that the multimodal explanations are consistent with human interpretations, help justify the models' decision, and provide useful insights to help diagnose an incorrect prediction. Finally, we describe a realworld ecommerce application for using the generated multimodal explanations.

ComputerAided Design as Language ; ComputerAided Design CAD applications are used in manufacturing to model everything from coffee mugs to sports cars. These programs are complex and require years of training and experience to master. A component of all CAD models particularly difficult to make are the highly structured 2D sketches that lie at the heart of every 3D construction. In this work, we propose a machine learning model capable of automatically generating such sketches. Through this, we pave the way for developing intelligent tools that would help engineers create better designs with less effort. Our method is a combination of a generalpurpose language modeling technique alongside an offtheshelf data serialization protocol. We show that our approach has enough flexibility to accommodate the complexity of the domain and performs well for both unconditional synthesis and imagetosketch translation.

Deviation from SlowRoll Regime in the EGB Inflationary Models with rsim Ne1 ; We consider EinsteinGaussBonnet EGB inflationary models using the effective potential approach. We present evolution equations in the slowroll regime using the effective potential and the tensortoscalar ratio. The choice of the effective potential is related to an expression of the spectral index in terms of efolding number Ne. The satisfaction of the slowroll regime is mostly related to the form of the tensortoscalar ratio r. The case of rsim1N2e leads to a generalization of alphaattractors inflationary parameters to EinsteinGaussBonnet gravity with exponential effective potential. Moreover, the cosmological attractors include models with rsim1Ne. And we check the satisfaction of the slowroll regime during inflation for models with rsim1Ne.

Autoencoder Based InterVehicle Generalization for InCabin Occupant Classification ; Common domain shift problem formulations consider the integration of multiple source domains, or the target domain during training. Regarding the generalization of machine learning models between different car interiors, we formulate the criterion of training in a single vehicle without access to the target distribution of the vehicle the model would be deployed to, neither with access to multiple vehicles during training. We performed an investigation on the SVIRO dataset for occupant classification on the rear bench and propose an autoencoder based approach to improve the transferability. The autoencoder is on par with commonly used classification models when trained from scratch and sometimes outperforms models pretrained on a large amount of data. Moreover, the autoencoder can transform images from unknown vehicles into the vehicle it was trained on. These results are corroborated by an evaluation on real infrared images from two vehicle interiors.

Generalising Multilingual ConcepttoText NLG with Language Agnostic Delexicalisation ; Concepttotext Natural Language Generation is the task of expressing an input meaning representation in natural language. Previous approaches in this task have been able to generalise to rare or unseen instances by relying on a delexicalisation of the input. However, this often requires that the input appears verbatim in the output text. This poses challenges in multilingual settings, where the task expands to generate the output text in multiple languages given the same input. In this paper, we explore the application of multilingual models in concepttotext and propose Language Agnostic Delexicalisation, a novel delexicalisation method that uses multilingual pretrained embeddings, and employs a characterlevel postediting model to inflect words in their correct form during relexicalisation. Our experiments across five datasets and five languages show that multilingual models outperform monolingual models in concepttotext and that our framework outperforms previous approaches, especially for low resource languages.

Neural Quality Estimation with Multiple Hypotheses for Grammatical Error Correction ; Grammatical Error Correction GEC aims to correct writing errors and help language learners improve their writing skills. However, existing GEC models tend to produce spurious corrections or fail to detect lots of errors. The quality estimation model is necessary to ensure learners get accurate GEC results and avoid misleading from poorly corrected sentences. Welltrained GEC models can generate several highquality hypotheses through decoding, such as beam search, which provide valuable GEC evidence and can be used to evaluate GEC quality. However, existing models neglect the possible GEC evidence from different hypotheses. This paper presents the Neural Verification Network VERNet for GEC quality estimation with multiple hypotheses. VERNet establishes interactions among hypotheses with a reasoning graph and conducts two kinds of attention mechanisms to propagate GEC evidence to verify the quality of generated hypotheses. Our experiments on four GEC datasets show that VERNet achieves stateoftheart grammatical error detection performance, achieves the best quality estimation results, and significantly improves GEC performance by reranking hypotheses. All data and source codes are available at httpsgithub.comthunlpVERNet.

Least squares Monte Carlo methods in stochastic Volterra rough volatility models ; In stochastic Volterra rough volatility models, the volatility follows a truncated Brownian semistationary process with stochastic volofvol. Recently, efficient VIX pricing Monte Carlo methods have been proposed for the case where the volofvol is Markovian and independent of the volatility. Following recent empirical data, we discuss the VIX option pricing problem for a generalized framework of these models, where the volofvol may depend on the volatility andor not be Markovian. In such a setting, the aforementioned Monte Carlo methods are not valid. Moreover, the classical least squares Monte Carlo faces exponentially increasing complexity with the number of grid time steps, whilst the nested Monte Carlo method requires a prohibitive number of simulations. By exploring the infinite dimensional Markovian representation of these models, we device a scalable least squares Monte Carlo for VIX option pricing. We apply our method firstly under the independence assumption for benchmarks, and then to the generalized framework. We also discuss the rough volofvol setting, where Markovianity of the volofvol is not present. We present simulations and benchmarks to establish the efficiency of our method.

Learning HighDimensional Distributions with Latent Neural FokkerPlanck Kernels ; Learning highdimensional distributions is an important yet challenging problem in machine learning with applications in various domains. In this paper, we introduce new techniques to formulate the problem as solving FokkerPlanck equation in a lowerdimensional latent space, aiming to mitigate challenges in highdimensional data space. Our proposed model consists of latentdistribution morphing, a generator and a parameterized FokkerPlanck kernel function. One fascinating property of our model is that it can be trained with arbitrary steps of latent distribution morphing or even without morphing, which makes it flexible and as efficient as Generative Adversarial Networks GANs. Furthermore, this property also makes our latentdistribution morphing an efficient plugandplay scheme, thus can be used to improve arbitrary GANs, and more interestingly, can effectively correct failure cases of the GAN models. Extensive experiments illustrate the advantages of our proposed method over existing models.

HuMoR 3D Human Motion Model for Robust Pose Estimation ; We introduce HuMoR a 3D Human Motion Model for Robust Estimation of temporal pose and shape. Though substantial progress has been made in estimating 3D human motion and shape from dynamic observations, recovering plausible pose sequences in the presence of noise and occlusions remains a challenge. For this purpose, we propose an expressive generative model in the form of a conditional variational autoencoder, which learns a distribution of the change in pose at each step of a motion sequence. Furthermore, we introduce a flexible optimizationbased approach that leverages HuMoR as a motion prior to robustly estimate plausible pose and shape from ambiguous observations. Through extensive evaluations, we demonstrate that our model generalizes to diverse motions and body shapes after training on a large motion capture dataset, and enables motion reconstruction from multiple input modalities including 3D keypoints and RGBD videos.

The Summary Loop Learning to Write Abstractive Summaries Without Examples ; This work presents a new approach to unsupervised abstractive summarization based on maximizing a combination of coverage and fluency for a given length constraint. It introduces a novel method that encourages the inclusion of key terms from the original document into the summary key terms are masked out of the original document and must be filled in by a coverage model using the current generated summary. A novel unsupervised training procedure leverages this coverage model along with a fluency model to generate and score summaries. When tested on popular news summarization datasets, the method outperforms previous unsupervised methods by more than 2 R1 points, and approaches results of competitive supervised methods. Our model attains higher levels of abstraction with copied passages roughly two times shorter than prior work, and learns to compress and merge sentences without supervision.

A general and fast convolutionbased method for peridynamics applications to elasticity and brittle fracture ; We introduce a general and fast convolutionbased method FCBM for peridynamics PD. Expressing the PD integrals in terms of convolutions and computing them by fast Fourier transform FFT, we reduce the computational complexity of PD models from ON2 to ONlog2 N, with N being the total number of discretization nodes. Initial neighbor identification and storing neighbor information is not required, and, as a consequence, memory allocation scales with ON instead of ON2, common for existing methods. The method is applicable to bounded domains with arbitrary shapes and boundary conditions via an embedded constraint EC approach. We explain the FCBMEC formulation for certain bondbased and statebased, linear and nonlinear PD models of elasticity and dynamic brittle fracture, as applications. We solve a 3D elastostatic problem and show that the FCBM reduces the computational time from days to hours and from years to days, compared with the original meshfree discretization for PD models. Largescale computations of PD models are feasible with the new method, and we demonstrate its versatility by simulating, with ease, the difficult problem of multiple crack branching in a brittle plate.

Counterfactual Interventions Reveal the Causal Effect of Relative Clause Representations on Agreement Prediction ; When language models process syntactically complex sentences, do they use their representations of syntax in a manner that is consistent with the grammar of the language We propose AlterRep, an interventionbased method to address this question. For any linguistic feature of a given sentence, AlterRep generates counterfactual representations by altering how the feature is encoded, while leaving intact all other aspects of the original representation. By measuring the change in a model's word prediction behavior when these counterfactual representations are substituted for the original ones, we can draw conclusions about the causal effect of the linguistic feature in question on the model's behavior. We apply this method to study how BERT models of different sizes process relative clauses RCs. We find that BERT variants use RC boundary information during word prediction in a manner that is consistent with the rules of English grammar; this RC boundary information generalizes to a considerable extent across different RC types, suggesting that BERT represents RCs as an abstract linguistic category.

Sentence Similarity Based on Contexts ; Existing methods to measure sentence similarity are faced with two challenges 1 labeled datasets are usually limited in size, making them insufficient to train supervised neural models; 2 there is a trainingtest gap for unsupervised language modeling LM based models to compute semantic scores between sentences, since sentencelevel semantics are not explicitly modeled at training. This results in inferior performances in this task. In this work, we propose a new framework to address these two issues. The proposed framework is based on the core idea that the meaning of a sentence should be defined by its contexts, and that sentence similarity can be measured by comparing the probabilities of generating two sentences given the same context. The proposed framework is able to generate highquality, largescale dataset with semantic similarity scores between two sentences in an unsupervised manner, with which the traintest gap can be largely bridged. Extensive experiments show that the proposed framework achieves significant performance boosts over existing baselines under both the supervised and unsupervised settings across different datasets.

Automating Cryptographic Protocol Language Generation from Structured Specifications ; Security of cryptographic protocols can be analysed by creating a model in a formal language and verifying the model in a tool. All such tools focus on the last part of the analysis, verification, and the interpretation of the specification is only explained in papers. Rather, we focus on the interpretation and modelling part by presenting a tool to aid the cryptographer throughout the process and automatically generating code in a target language. We adopt a datacentric approach where the protocol design is stored in a structured way rather than as textual specifications. Previous work shows how this approach facilitates the interpretation to a single language for Tamarin which required aftermath modifications. By improving the expressiveness of the specification data structure we extend the tool to export to an additional formal language, ProVerif, as well as a C fully running implementation. Furthermore, we extend the plugins to verify correctness in ProVerif and executability lemmas in Tamarin. In this paper we model the DiffieHellman key exchange, which is traditionally used as a case study; a demo is also provided for other commonly studied protocols, Needham Schroeder and NeedhamSchroederLowe.

Vector autoregression models with skewness and heavy tails ; With uncertain changes of the economic environment, macroeconomic downturns during recessions and crises can hardly be explained by a Gaussian structural shock. There is evidence that the distribution of macroeconomic variables is skewed and heavy tailed. In this paper, we contribute to the literature by extending a vector autoregression VAR model to account for a more realistic assumption of the multivariate distribution of the macroeconomic variables. We propose a general class of generalized hyperbolic skew Student's t distribution with stochastic volatility for the error term in the VAR model that allows us to take into account skewness and heavy tails. Tools for Bayesian inference and model selection using a Gibbs sampler are provided. In an empirical study, we present evidence of skewness and heavy tails for monthly macroeconomic variables. The analysis also gives a clear message that skewness should be taken into account for better predictions during recessions and crises.

Sampling random graphs with specified degree sequences ; The configuration model is a standard tool for uniformly generating random graphs with a specified degree sequence, and is often used as a null model to evaluate how much of an observed network's structure can be explained by its degree structure alone. A Markov chain Monte Carlo MCMC algorithm, based on a degreepreserving doubleedge swap, provides an asymptotic solution to sample from the configuration model. However, accurately and efficiently detecting this Markov chain's convergence on its stationary distribution remains an unsolved problem. Here, we provide a solution to detect convergence and sample from the configuration model. We develop an algorithm, based on the assortativity of the sampled graphs, for estimating the gap between effectively independent MCMC states, and a computationally efficient gapestimation heuristic derived from analyzing a corpus of 509 empirical networks. We provide a convergence detection method based on the DickeyFuller Generalized Least Squares test, which we show is more accurate and efficient than three alternative Markov chain convergence tests.

Language Model as an Annotator Exploring DialoGPT for Dialogue Summarization ; Current dialogue summarization systems usually encode the text with a number of general semantic features e.g., keywords and topics to gain more powerful dialogue modeling capabilities. However, these features are obtained via opendomain toolkits that are dialogagnostic or heavily relied on human annotations. In this paper, we show how DialoGPT, a pretrained model for conversational response generation, can be developed as an unsupervised dialogue annotator, which takes advantage of dialogue background knowledge encoded in DialoGPT. We apply DialoGPT to label three types of features on two dialogue summarization datasets, SAMSum and AMI, and employ pretrained and non pretrained models as our summarizes. Experimental results show that our proposed method can obtain remarkable improvements on both datasets and achieves new stateoftheart performance on the SAMSum dataset.

Efficient HighResolution ImagetoImage Translation using MultiScale Gradient UNet ; Recently, Conditional Generative Adversarial Network Conditional GAN have shown very promising performance in several imagetoimage translation applications. However, the uses of these conditional GANs are quite limited to lowresolution images, such as 256X256.The Pix2PixHD is a recent attempt to utilize the conditional GAN for highresolution image synthesis. In this paper, we propose a MultiScale Gradient based UNet MSG UNet model for highresolution imagetoimage translation up to 2048X1024 resolution. The proposed model is trained by allowing the flow of gradients from multiplediscriminators to a single generator at multiple scales. The proposed MSG UNet architecture leads to photorealistic highresolution imagetoimage translation. Moreover, the proposed model is computationally efficient as compared to the Pix2PixHD with an improvement in the inference time nearly by 2.5 times. We provide the code of MSG UNet model at httpsgithub.comlaxmanironMSGUNet.

Support vector machines and linear regression coincide with very highdimensional features ; The support vector machine SVM and minimum Euclidean norm least squares regression are two fundamentally different approaches to fitting linear models, but they have recently been connected in models for very highdimensional data through a phenomenon of support vector proliferation, where every training example used to fit an SVM becomes a support vector. In this paper, we explore the generality of this phenomenon and make the following contributions. First, we prove a superlinear lower bound on the dimension in terms of sample size required for support vector proliferation in independent feature models, matching the upper bounds from previous works. We further identify a sharp phase transition in Gaussian feature models, bound the width of this transition, and give experimental support for its universality. Finally, we hypothesize that this phase transition occurs only in much higherdimensional settings in the ell1 variant of the SVM, and we present a new geometric characterization of the problem that may elucidate this phenomenon for the general ellp case.

Strange star with KroriBarua potential in presence of anisotropy ; In present paper a wellbehaved new model of anisotropic compact star in 31dimensional spacetime has been investigated in the background of Einstein's general theory of relativity. The model has been developed by choosing grr component as KroriBarua KB ansatz Krori and Barua in J. Phys. A, Math. Gen. 8508, 1975. The field equations have been solved by a proper choice of the anisotropy factor which is physically reasonable and well behaved inside the stellar interior. Interior spacetime has been matched smoothly to the exterior Schwarzschild vacuum solution and it has also been depicted graphically. Model is free from all types of singularities and is in static equilibrium under different forces acting on the system. The stability of the model has been tested with the help of various conditions available in literature. The solution is compatible with observed masses and radii of a few compact stars.

Reinforce Security A ModelFree Approach Towards Secure Wiretap Coding ; The use of deep learningbased techniques for approximating secure encoding functions has attracted considerable interest in wireless communications due to impressive results obtained for general coding and decoding tasks for wireless communication systems. Of particular importance is the development of modelfree techniques that work without knowledge about the underlying channel. Such techniques utilize for example generative adversarial networks to estimate and model the conditional channel distribution, mutual information estimation as a reward function, or reinforcement learning. In this paper, the approach of reinforcement learning is studied and, in particular, the policy gradient method for a modelfree approach of neural networkbased secure encoding is investigated. Previously developed techniques for enforcing a certain coset structure on the encoding process can be combined with recent reinforcement learning approaches. This new approach is evaluated by extensive simulations, and it is demonstrated that the resulting decoding performance of an eavesdropper is capped at a certain error level.

A Computer Program for the Numerical Analysis of Economic Cycles Within the Framework of the Dubovsky Generalized Model ; The article proposes a computer program for calculating economic crises according to the generalized mathematical model of S.V. Dubovsky. This model is represented by a system of ordinary nonlinear differential equations with fractional derivatives in the sense of GerasimovCaputo with initial conditions. Furthermore, according to a numerical algorithm based on an explicit nonlocal finitedifference scheme, oscillograms and phase trajectories were constructed. It is shown that changing the orders of fractional derivatives in the model can give rise to various modes, for example, damped modes with a steadystate amplitude. It is concluded that the orders of fractional derivatives are responsible for the intensity of the process.

The Image Local Autoregressive Transformer ; Recently, AutoRegressive AR models for the whole image generation empowered by transformers have achieved comparable or even better performance to Generative Adversarial Networks GANs. Unfortunately, directly applying such AR models to editchange local image regions, may suffer from the problems of missing global information, slow inference speed, and information leakage of local guidance. To address these limitations, we propose a novel model image Local Autoregressive Transformer iLAT, to better facilitate the locally guided image synthesis. Our iLAT learns the novel local discrete representations, by the newly proposed local autoregressive LA transformer of the attention mask and convolution mechanism. Thus iLAT can efficiently synthesize the local image regions by key guidance information. Our iLAT is evaluated on various locally guided image syntheses, such as poseguided person image synthesis and face editing. Both the quantitative and qualitative results show the efficacy of our model.

Neural semiMarkov CRF for Monolingual Word Alignment ; Monolingual word alignment is important for studying finegrained editing operations i.e., deletion, addition, and substitution in texttotext generation tasks, such as paraphrase generation, text simplification, neutralizing biased language, etc. In this paper, we present a novel neural semiMarkov CRF alignment model, which unifies word and phrase alignments through variablelength spans. We also create a new benchmark with human annotations that cover four different text genres to evaluate monolingual word alignment models in more realistic settings. Experimental results show that our proposed model outperforms all previous approaches for monolingual word alignment as well as a competitive QAbased baseline, which was previously only applied to bilingual data. Our model demonstrates good generalizability to three outofdomain datasets and shows great utility in two downstream applications automatic text simplification and sentence pair classification tasks.

Evaluation of cosmological models in fR, T gravity in different dark energy scenario ; In present paper, we search the existence of dark energy scalar field models within in fR, T gravity theory established by Harko et al. Phys. Rev. D 84, 024020, 2011 in a flat FRW universe. The correspondence between scalar field models have been examined by employing new generalized dynamical cosmological term Lambdat . In this regards, the best fit observational values of parameters from three distinct sets data are applied. To decide the solution to field equations, a scale factor a leftsinhbeta tright1n has been considered, where beta n are constants. Here, we employ the recent ensues H069.2 and q00.52 from OHDJLA observation Yu et al., Astrophys. J. 856, 3, 2018. Through the numerical estimation and graphical assessing of various cosmological parameters, it has been experienced that findings are comparable with kinematics and physical properties of universe and compatible with recent cosmological ensues. The dynamics and potentials of scalar fields are clarified in FRW scenario in the present model. Potentials reconstruction is highly reasonable and shows a periodic establishment and in agreement with latest observations.

Baseline Skinning for Point Sets of Articulated Bodies ; General skinning techniques aim to deform the surface of an articulated model following the pose change of a skeleton. Their rapidity makes them ideal tools for realtime animation purposes. However, popular skinning algorithms are simple, but they tend to generate undesirable geometric artefacts. In our work, we consider skeletons given in the form of spheremesh models controlling both the pose and morphology of the shape that is either described as a mesh or a raw point set. We propose a novel skinning method that encodes the point set details above a bundle of baselines covering the spheremesh. In particular, we propose a geometrical model of the baseline and detail direction evolution during bone twisting and joints bending rotations. Our approach works directly on point sets and thus preserves the accuracy of the initial sampling. It further avoids computing a weight per point or a costly explicit muscle modelling step. We evaluate our method on several articulated body point sets, showing that it creates fewer artefacts than classical methods.

Model Predictive RobotEnvironment Interaction Control for Mobile Manipulation Tasks ; Modern, torquecontrolled service robots can regulate contact forces when interacting with their environment. Model Predictive Control MPC is a powerful method to solve the underlying control problem, allowing to plan for wholebody motions while including different constraints imposed by the robot dynamics or its environment. However, an accurate model of the robotenvironment is needed to achieve a satisfying closedloop performance. Currently, this necessity undermines the performance and generality of MPC in manipulation tasks. In this work, we combine an MPCbased wholebody controller with two adaptive schemes, derived from online system identification and adaptive control. As a result, we enable a general mobile manipulator to interact with unknown environments, without any need for retuning parameters or premodeling the interacting objects. In combination with the MPC controller, the two adaptive approaches are validated and benchmarked with a ballbalancing manipulator in door opening and object lifting tasks.

Bayesian Attention Belief Networks ; Attentionbased neural networks have achieved stateoftheart results on a wide range of tasks. Most such models use deterministic attention while stochastic attention is less explored due to the optimization difficulties or complicated model design. This paper introduces Bayesian attention belief networks, which construct a decoder network by modeling unnormalized attention weights with a hierarchy of gamma distributions, and an encoder network by stacking Weibull distributions with a deterministicupwardstochasticdownward structure to approximate the posterior. The resulting autoencoding networks can be optimized in a differentiable way with a variational lower bound. It is simple to convert any models with deterministic attention, including pretrained ones, to the proposed Bayesian attention belief networks. On a variety of language understanding tasks, we show that our method outperforms deterministic attention and stateoftheart stochastic attention in accuracy, uncertainty estimation, generalization across domains, and robustness to adversarial attacks. We further demonstrate the general applicability of our method on neural machine translation and visual question answering, showing great potential of incorporating our method into various attentionrelated tasks.

Probing transfer learning with a model of synthetic correlated datasets ; Transfer learning can significantly improve the sample efficiency of neural networks, by exploiting the relatedness between a datascarce target task and a dataabundant source task. Despite years of successful applications, transfer learning practice often relies on adhoc solutions, while theoretical understanding of these procedures is still limited. In the present work, we rethink a solvable model of synthetic data as a framework for modeling correlation between datasets. This setup allows for an analytic characterization of the generalization performance obtained when transferring the learned feature map from the source to the target task. Focusing on the problem of training twolayer networks in a binary classification setting, we show that our model can capture a range of salient features of transfer learning with real data. Moreover, by exploiting parametric control over the correlation between the two datasets, we systematically investigate under which conditions the transfer of features is beneficial for generalization.

Learning to See by Looking at Noise ; Current vision systems are trained on huge datasets, and these datasets come with costs curation is expensive, they inherit human biases, and there are concerns over privacy and usage rights. To counter these costs, interest has surged in learning from cheaper data sources, such as unlabeled images. In this paper we go a step further and ask if we can do away with real image datasets entirely, instead learning from noise processes. We investigate a suite of image generation models that produce images from simple random processes. These are then used as training data for a visual representation learner with a contrastive loss. We study two types of noise processes, statistical image models and deep generative models under different random initializations. Our findings show that it is important for the noise to capture certain structural properties of real data but that good performance can be achieved even with processes that are far from realistic. We also find that diversity is a key property to learn good representations. Datasets, models, and code are available at httpsmbaradad.github.iolearningwithnoise.

Deep Probabilistic Koopman Longterm timeseries forecasting under periodic uncertainties ; Probabilistic forecasting of complex phenomena is paramount to various scientific disciplines and applications. Despite the generality and importance of the problem, general mathematical techniques that allow for stable longterm forecasts with calibrated uncertainty measures are lacking. For most time series models, the difficulty of obtaining accurate probabilistic future time step predictions increases with the prediction horizon. In this paper, we introduce a surprisingly simple approach that characterizes timevarying distributions and enables reasonably accurate predictions thousands of timesteps into the future. This technique, which we call Deep Probabilistic Koopman DPK, is based on recent advances in linear Koopman operator theory, and does not require time stepping for future time predictions. Koopman models also tend to have a small parameter footprint often less than 10,000 parameters. We demonstrate the longterm forecasting performance of these models on a diversity of domains, including electricity demand forecasting, atmospheric chemistry, and neuroscience. For electricity demand modeling, our domainagnostic technique outperforms all of 177 domainspecific competitors in the most recent Global Energy Forecasting Competition.

Conditional Variational Autoencoder with Adversarial Learning for EndtoEnd TexttoSpeech ; Several recent endtoend texttospeech TTS models enabling singlestage training and parallel sampling have been proposed, but their sample quality does not match that of twostage TTS systems. In this work, we present a parallel endtoend TTS method that generates more natural sounding audio than current twostage models. Our method adopts variational inference augmented with normalizing flows and an adversarial training process, which improves the expressive power of generative modeling. We also propose a stochastic duration predictor to synthesize speech with diverse rhythms from input text. With the uncertainty modeling over latent variables and the stochastic duration predictor, our method expresses the natural onetomany relationship in which a text input can be spoken in multiple ways with different pitches and rhythms. A subjective human evaluation mean opinion score, or MOS on the LJ Speech, a single speaker dataset, shows that our method outperforms the best publicly available TTS systems and achieves a MOS comparable to ground truth.

Piecewiseconstant Neural ODEs ; Neural networks are a popular tool for modeling sequential data but they generally do not treat time as a continuous variable. Neural ODEs represent an important exception they parameterize the time derivative of a hidden state with a neural network and then integrate over arbitrary amounts of time. But these parameterizations, which have arbitrary curvature, can be hard to integrate and thus train and evaluate. In this paper, we propose making a piecewiseconstant approximation to Neural ODEs to mitigate these issues. Our model can be integrated exactly via Euler integration and can generate autoregressive samples in 320 times fewer steps than comparable RNN and ODERNN models. We evaluate our model on several synthetic physics tasks and a planning task inspired by the game of billiards. We find that it matches the performance of baseline approaches while requiring less time to train and evaluate.

Correlated NonCoherent Radar Detection for GammaFluctuating Targets in Compound Clutter ; This work studies the problem of radar detection of correlated gammafluctuating targets in the presence of clutter described by compound models with correlated speckle. If the correlation is not accounted for in a radar model, the required signaltointerference ratio for a given probability of detection will be incorrect, resulting in overestimated performance. Although more generally applicable, the is focus on airborne maritime radar systems. Hence Kdistributed sea clutter is used as the main example. Detection via squarelaw noncoherent pulse integration is formulated in a way that accommodates arbitrary partial correlation for both target radar crosssection RCS and clutter speckle. The obstacle to including this degree of generality in previous work was the fact that Swerling's original characterization of the standard RCS fluctuation classes as gamma distributions for the power is not sufficient for the inclusion of both correlation sources i.e.target and clutter speckle for gammafluctuating targets. An extension of the model is required at the quadrature component i.e. voltage level, as phase relationships can no longer be neglected. This is addressed in the present work, which not only postulates an extended model, but also demonstrates how to efficiently compute it, with and without a number of simplifying approximation schemes within the framework of the saddlepoint technique.

Environment dependent vibrational heat transport in molecular Junctions Rectification, quantum effects, vibrational mismatch ; Vibrational heat transport in molecular junctions is a central issue in different contemporary research areas like Chemistry, material science, mechanical engineering, thermoelectrics and power generation. Our model system consists of a chain of molecules which sandwiched between two solids that are maintained at different temperatures. We employ quantum selfconsistent reservoir model, which is built on generalized quantum Langevin equation, to investigate quantum effects and far from equilibrium conditions on thermal conduction at nanoscale. The present selfconsistent reservoir model can easily mimic the phononphonon scattering mechanisms. Different thermal environments are modelled as i Ohmic, ii subOhmic, and iii superOhmic environment and their effects are demonstrated for the thermal rectification properties of the system with spring graded or mass graded feature. The behavior of heat current across molecular junctions as a function of chain length, temperature gradient and phonon scattering rate are studied. Further, our analysis reveals the effects of vibrational mismatch between the solids phonon spectra on heat transfer characteristics in molecular junctions for different thermal environments.

Implementing Permutations in the Brain and SVO Frequencies of Languages ; The subjectverbobject SVO word order prevalent in English is shared by about 42 of world languages. Another 45 of all languages follow the SOV order, 9 the VSO order, and fewer languages use the three remaining permutations. None of the many extant explanations of this phenomenon take into account the difficulty of implementing these permutations in the brain. We propose a plausible model of sentence generation inspired by the recently proposed Assembly Calculus framework of brain function. Our model results in a natural explanation of the uneven frequencies. Estimating the parameters of this model yields predictions of the relative difficulty of disinhibiting one brain area from another. Our model is based on the standard syntax tree, a simple binary tree with three leaves. Each leaf corresponds to one of the three parts of a basic sentence. The leaves can be activated through lock and unlock operations and the sequence of activation of the leaves implements a specific word order. More generally, we also formulate and algorithmically solve the problems of implementing a permutation of the leaves of any binary tree, and of selecting the permutation that is easiest to implement on a given binary tree.

A Supersymmetric Flavor Clockwork ; Clockwork models can explain the flavor hierarchies in the Standard Model quark and lepton spectrum. We construct supersymmetric versions of such flavor clockwork models. The zero modes of the clockwork are identified with the fermions and sfermions of the Minimal Supersymmetric Standard Model. In addition to generating a hierarchical fermion spectrum, the clockwork also predicts a specific flavor structure for the soft SUSY breaking sfermion masses. We find sizeable flavor mixing among first and second generation squarks. Constraints from Kaon oscillations require the masses of either squarks or gluinos to be above a scale of sim 3 PeV.

Multiphase turbulence modeling using sparse regression and gene expression programming ; In recent years, there has been an explosion of machine learning techniques for turbulence closure modeling, though many rely on augmenting existing models. While this has proven successful in singlephase flows, it breaks down for multiphase flows, particularly when the system dynamics are controlled by twoway coupling between the phases. In this work, we propose an approach that blends sparse regression and gene expression programming GEP to generate closedform algebraic models from simulation data. Sparse regression is used to determine a minimum set of functional groups required to capture the physics and GEP is used to automate the formulation of the coefficients and dependencies on operating conditions. The framework is demonstrated on a canonical gassolid flow in which twoway coupling generates and sustains fluidphase turbulence.

On the bimodal Gumbel model with application to environmental data ; The Gumbel model is a very popular statistical model due to its wide applicability for instance in the course of certain survival, environmental, financial or reliability studies. In this work, we have introduced a bimodal generalization of the Gumbel distribution that can be an alternative to model bimodal data. We derive the analytical shapes of the corresponding probability density function and the hazard rate function and provide graphical illustrations. Furthermore, We have discussed the properties of this density such as mode, bimodality, moment generating function and moments. Our results were verified using the Markov chain Monte Carlo simulation method. The maximum likelihood method is used for parameters estimation. Finally, we also carry out an application to real data that demonstrates the usefulness of the proposed distribution.

Relativistic compact stars in Tolman spacetime via an anisotropic approach ; In this present work, we have obtained a singularityfree spherically symmetric stellar model with anisotropic pressure in the background of Einstein's general theory of relativity. The Einstein's field equations have been solved by exploiting Tolman em ansatz Richard C Tolman, Phys. Rev. 55364, 1939 in 31dimensional spacetime. Using observed values of mass and radius of the compact star PSR J1903327, we have calculated the numerical values of all the constants from the boundary conditions. All the physical characteristics of the proposed model have been discussed both analytically and graphically. The new exact solution satisfies all the physical criteria for a realistic compact star. The matter variables are regular and well behaved throughout the stellar structure. Constraints on model parameters have been obtained. All the energy conditions are verified with the help of graphical representation. The stability condition of the present model has been described through different testings.

JointGT GraphText Joint Representation Learning for Text Generation from Knowledge Graphs ; Existing pretrained models for knowledgegraphtotext KGtotext generation simply finetune texttotext pretrained models such as BART or T5 on KGtotext datasets, which largely ignore the graph structure during encoding and lack elaborate pretraining tasks to explicitly model graphtext alignments. To tackle these problems, we propose a graphtext joint representation learning model called JointGT. During encoding, we devise a structureaware semantic aggregation module which is plugged into each Transformer layer to preserve the graph structure. Furthermore, we propose three new pretraining tasks to explicitly enhance the graphtext alignment including respective text graph reconstruction, and graphtext alignment in the embedding space via Optimal Transport. Experiments show that JointGT obtains new stateoftheart performance on various KGtotext datasets.

Ensemble of ACCDOA and EINV2based Systems with D3Nets and Impulse Response Simulation for Sound Event Localization and Detection ; This report describes our systems submitted to the DCASE2021 challenge task 3 sound event localization and detection SELD with directional interference. Our previous system based on activitycoupled Cartesian direction of arrival ACCDOA representation enables us to solve a SELD task with a single target. This ACCDOAbased system with efficient network architecture called RD3Net and data augmentation techniques outperformed stateoftheart SELD systems in terms of localization and locationdependent detection. Using the ACCDOAbased system as a base, we perform model ensembles by averaging outputs of several systems trained with different conditions such as input features, training folds, and model architectures. We also use the event independent network v2 EINV2based system to increase the diversity of the model ensembles. To generalize the models, we further propose impulse response simulation IRS, which generates simulated multichannel signals by convolving simulated room impulse responses RIRs with source signals extracted from the original dataset. Our systems significantly improved over the baseline system on the development dataset.

An Approach to Causal Inference over Stochastic Networks ; Claiming causal inferences in network settings necessitates careful consideration of the often complex dependency between outcomes for actors. Of particular importance are treatment spillover or outcome interference effects. We consider causal inference when the actors are connected via an underlying network structure. Our key contribution is a model for causality when the underlying network is unobserved and the actor covariates evolve stochastically over time. We develop a joint model for the relational and covariate generating process that avoids restrictive separability assumptions and deterministic network assumptions that do not hold in the majority of social network settings of interest. Our framework utilizes the highly general class of Exponentialfamily Random Network models ERNM of which Markov Random Fields MRF and Exponentialfamily Random Graph models ERGM are special cases. We present potential outcome based inference within a Bayesian framework, and propose a simple modification to the exchange algorithm to allow for sampling from ERNM posteriors. We present results of a simulation study demonstrating the validity of the approach. Finally, we demonstrate the value of the framework in a casestudy of smoking over time in the context of adolescent friendship networks.

Graphical Economies with Resale ; Kakade, Kearns, and Ortiz KKO introduce a graphtheoretic generalization of the classic ArrowDebreu AD exchange economy. Despite its appeal as a networked version of AD, we argue that the KKO model is too local, in the sense that goods cannot travel more than one hop through the network. We introduce an alternative model in which agents may purchase goods on credit in order to resell them. In contrast to KKO, our model allows for longrange trade, and yields equilibria in more settings than KKO, including sparse endowments. Our model smoothly interpolates between the KKO and AD equilibrium concepts we recover KKO when the resale capacity is zero, and recover AD when it is sufficiently large. We give general equilibrium existence results, and an auctionbased algorithm to compute approximate equilibria when agent utilities satisfy the weak grosssubstitutes property.

GraphAnoGAN Detecting Anomalous Snapshots from Attributed Graphs ; Finding anomalous snapshots from a graph has garnered huge attention recently. Existing studies address the problem using shallow learning mechanisms such as subspace selection, egonetwork, or community analysis. These models do not take into account the multifaceted interactions between the structure and attributes in the network. In this paper, we propose GraphAnoGAN, an anomalous snapshot ranking framework, which consists of two core components generative and discriminative models. Specifically, the generative model learns to approximate the distribution of anomalous samples from the candidate set of graph snapshots, and the discriminative model detects whether the sampled snapshot is from the groundtruth or not. Experiments on 4 realworld networks show that GraphAnoGAN outperforms 6 baselines with a significant margin 28.29 and 22.01 higher precision and recall, respectively compared to the best baseline, averaged across all datasets.

AdaPTGMM Powerful and robust covariateassisted multiple testing ; We propose a new empirical Bayes method for covariateassisted multiple testing with false discovery rate FDR control, where we model the local false discovery rate for each hypothesis as a function of both its covariates and pvalue. Our method refines the adaptive pvalue thresholding AdaPT procedure by generalizing its masking scheme to reduce the bias and variance of its false discovery proportion estimator, improving the power when the rejection set is small or some null pvalues concentrate near 1. We also introduce a Gaussian mixture model for the conditional distribution of the test statistics given covariates, modeling the mixing proportions with a generic userspecified classifier, which we implement using a twolayer neural network. Like AdaPT, our method provably controls the FDR in finite samples even if the classifier or the Gaussian mixture model is misspecified. We show in extensive simulations and real data examples that our new method, which we call AdaPTGMM, consistently delivers high power relative to competing stateoftheart methods. In particular, it performs well in scenarios where AdaPT is underpowered, and is especially wellsuited for testing composite null hypothesis, such as whether the effect size exceeds a practical significance threshold.

Long Shortterm Cognitive Networks ; In this paper, we present a recurrent neural system named Long Shortterm Cognitive Networks LSTCNs as a generalization of the Shortterm Cognitive Network STCN model. Such a generalization is motivated by the difficulty of forecasting very long time series efficiently. The LSTCN model can be defined as a collection of STCN blocks, each processing a specific time patch of the multivariate time series being modeled. In this neural ensemble, each block passes information to the subsequent one in the form of weight matrices representing the prior knowledge. As a second contribution, we propose a deterministic learning algorithm to compute the learnable weights while preserving the prior knowledge resulting from previous learning processes. As a third contribution, we introduce a feature influence score as a proxy to explain the forecasting process in multivariate time series. The simulations using three case studies show that our neural system reports small forecasting errors while being significantly faster than stateoftheart recurrent models.

A simple parity violating model in the symmetric teleparallel gravity and its cosmological perturbations ; The parity violating gravity models based on the symmetric teleparallel gravity have been considered in the literature, but their applications in cosmology and especially the modifications to cosmological perturbations have not been fully explored. In this paper we consider such a simple model which modifies general relativity by a parity nonconserved coupling, within the framework of the symmetric teleparallel gravity. We study in detail its cosmological applications and focus on its cosmological perturbation theory. Besides the already known parity violation in the tensor perturbations, we find that the vector perturbations in this model are promoted to be dynamical degrees of freedom, and the left and righthanded vector modes propagate with different velocities. More importantly, we find that one of the vector modes is a ghost at high momentum scales, which will give rise to the problem of vacuum instability in the quantum theory of cosmological perturbations.

Challenges in identifying simple patternforming mechanisms in the development of settlements using demographic data ; The rapid increase of population and settlement structures in the Global South during recent decades motivates the development of suitable models to describe their formation and evolution. Such settlement formation has been previously suggested to be dynamically driven by simple patternforming mechanisms. Here, we explore the use of a datadriven whitebox approach, called SINDy, to discover differential equation models directly from available spatiotemporal demographic data for three representative regions of the Global South. We show that the current resolution and observation time of the available data is insufficient to uncover relevant patternforming mechanisms in settlement development. Using synthetic data generated with a generic patternforming model, the AllenCahn equation, we characterize what the requirements are on spatial and temporal resolution, as well as observation time, to successfully identify possible model system equations. Overall, the study provides a theoretical framework for the analysis of largescale geographicalecological systems, and it motivates further improvements in optimization approaches and data collection.

Improving Counterfactual Generation for Fair Hate Speech Detection ; Bias mitigation approaches reduce models' dependence on sensitive features of data, such as social group tokens SGTs, resulting in equal predictions across the sensitive features. In hate speech detection, however, equalizing model predictions may ignore important differences among targeted social groups, as hate speech can contain stereotypical language specific to each SGT. Here, to take the specific language about each SGT into account, we rely on counterfactual fairness and equalize predictions among counterfactuals, generated by changing the SGTs. Our method evaluates the similarity in sentence likelihoods via pretrained language models among counterfactuals, to treat SGTs equally only within interchangeable contexts. By applying logit pairing to equalize outcomes on the restricted set of counterfactuals for each instance, we improve fairness metrics while preserving model performance on hate speech detection.

Finetuning GPT3 for Russian Text Summarization ; Automatic summarization techniques aim to shorten and generalize information given in the text while preserving its core message and the most relevant ideas. This task can be approached and treated with a variety of methods, however, not many attempts have been made to produce solutions specifically for the Russian language despite existing localizations of the stateoftheart models. In this paper, we aim to showcase ruGPT3 ability to summarize texts, finetuning it on the corpora of Russian news with their corresponding humangenerated summaries. Additionally, we employ hyperparameter tuning so that the model's output becomes less random and more tied to the original text. We evaluate the resulting texts with a set of metrics, showing that our solution can surpass the stateoftheart model's performance without additional changes in architecture or loss function. Despite being able to produce sensible summaries, our model still suffers from a number of flaws, namely, it is prone to altering Named Entities present in the original text such as surnames, places, dates, deviating from facts stated in the given document, and repeating the information in the summary.

Conditional Temporal Variational AutoEncoder for Action Video Prediction ; To synthesize a realistic action sequence based on a single human image, it is crucial to model both motion patterns and diversity in the action video. This paper proposes an Action Conditional Temporal Variational AutoEncoder ACTVAE to improve motion prediction accuracy and capture movement diversity. ACTVAE predicts pose sequences for an action clips from a single input image. It is implemented as a deep generative model that maintains temporal coherence according to the action category with a novel temporal modeling on latent space. Further, ACTVAE is a general action sequence prediction framework. When connected with a plugandplay PosetoImage P2I network, ACTVAE can synthesize image sequences. Extensive experiments bear out our approach can predict accurate pose and synthesize realistic image sequences, surpassing stateoftheart approaches. Compared to existing methods, ACTVAE improves model accuracy and preserves diversity.

Yoneda Lemma for mathcalDSimplicial Spaces ; For a small category mathcalD we define fibrations of simplicial presheaves on the category mathcalDtimesDelta, which we call localized mathcalDleft fibration. We show these fibrations can be seen as fibrant objects in a model structure, the localized mathcalDcovariant model structure, that is Quillen equivalent to a category of functors valued in simplicial presheaves on mathcalD, where the Quillen equivalence is given via a generalization of the Grothendieck construction. We use our understanding of this construction to give a detailed characterization of fibrations and weak equivalences in this model structure and in particular obtain a Yoneda lemma. We apply this general framework to study Cartesian fibrations of infty,ncategories, for models of infty,ncategories that arise via simplicial presheaves, such as nfold complete Segal spaces. This, in particular, results in the Yoneda lemma and Grothendieck construction for Cartesian fibrations of infty,ncategories.

BoseEinstein condensation processes with nontrivial geometric multiplicites realized via cal PTsymmetric and exactly solvable linearBoseHubbard building blocks ; It is well known that using the conventional nonHermitian but cal PTsymmetric BoseHubbard Hamiltonian with real spectrum one can realize the BoseEinstein condensation BEC process in an exceptionalpoint limit of order N. Such an exactly solvable simulation of the BECtype phase transition is, unfortunately, incomplete because the standard version of the model only offers an extreme form of the limit characterized by a minimal geometric multiplicity K1. In our paper we describe a rescaled and partitioned directsum modification of the linear version of the BoseHubbard model which remains exactly solvable while admitting any value of Kgeq 1. It offers a complete menu of benchmark models numbered by a specific combinatorial scheme. In this manner, an exhaustive classification of the general BEC patterns with any geometric multiplicity is obtained and realized in terms of an exactly solvable generalized BoseHubbard model.

UncertaintyAware Model Adaptation for Unsupervised CrossDomain Object Detection ; This work tackles the unsupervised crossdomain object detection problem which aims to generalize a pretrained object detector to a new target domain without labels. We propose an uncertaintyaware model adaptation method, which is based on two motivations 1 the estimation and exploitation of model uncertainty in a new domain is critical for reliable domain adaptation; and 2 the joint alignment of distributions for inputs feature alignment and outputs selftraining is needed. To this end, we compose a Bayesian CNNbased framework for uncertainty estimation in object detection, and propose an algorithm for generation of uncertaintyaware pseudolabels. We also devise a scheme for joint feature alignment and selftraining of the object detection model with uncertaintyaware pseudolabels. Experiments on multiple crossdomain object detection benchmarks show that our proposed method achieves stateoftheart performance.

A Mechanically Verified Theory of Contracts ; Cyberphysical systems CPS are assemblies of networked, heterogeneous, hardware, and software components sensing, evaluating, and actuating a physical environment. This heterogeneity induces complexity that makes CPSs challenging to model correctly. Since CPSs often have critical functions, it is however of utmost importance to formally verify them in order to provide the highest guarantees of safety. Faced with CPS complexity, model abstraction becomes paramount to make verification attainable. To this end, assumeguarantee contracts enable component model abstraction to support a sound, structured, and modular verification process. While abstractions of models by contracts are usually proved sound, none of the related contract frameworks themselves have, to the best of our knowledge, been formally proved correct so far. In this aim, we present the formalization of a generic assumeguarantee contract theory in the proof assistant Coq. We identify and prove theorems that ensure its correctness. Our theory is generic, or parametric, in that it can be instantiated and used with any given logic, in particular hybrid logics, in which highly complex cyberphysical systems can uniformly be described.

On Generalized Random Environment INAR Models of Higher Order Estimation of Random Environment States ; The behavior of a generalized random environment integervalued autoregressive model of higher order with geometric marginal distribution and negative binomial thinning operator abbrev. RrNGINARmathcalM,A,P is dictated by a realization znn1infty of an auxiliary Markov chain called random environment process. Element zn represents a state of the environment in moment ninmathbbN and determines three different parameters of the model in that moment. In order to use RrNGINARmathcalM,A,P model, one first needs to estimate znn1infty, which was so far done by Kmeans data clustering. We argue that this approach ignores some information and performs poorly in certain situations. We propose a new method for estimating znn1infty, which includes the data transformation preceding the clustering, in order to reduce the information loss. To confirm its efficiency, we compare this new approach with the usual one when applied on the simulated and the reallife data, and notice all the benefits obtained from our method.

Multitask learning from fixedwing UAV images for 2D3D city modeling ; Singletask learning in artificial neural networks will be able to learn the model very well, and the benefits brought by transferring knowledge thus become limited. In this regard, when the number of tasks increases e.g., semantic segmentation, panoptic segmentation, monocular depth estimation, and 3D point cloud, duplicate information may exist across tasks, and the improvement becomes less significant. Multitask learning has emerged as a solution to knowledgetransfer issues and is an approach to scene understanding which involves multiple related tasks each with potentially limited training data. Multitask learning improves generalization by leveraging the domainspecific information contained in the training data of related tasks. In urban management applications such as infrastructure development, traffic monitoring, smart 3D cities, and change detection, automated multitask data analysis for scene understanding based on the semantic, instance, and panoptic annotation, as well as monocular depth estimation, is required to generate precise urban models. In this study, a common framework for the performance assessment of multitask learning methods from fixedwing UAV images for 2D3D city modeling is presented.

Scalar field models driven by DiracBornInfeld dynamics and their relatives ; In this paper, we investigate novel kinklike structures in a scalar field theory driven by DiracBornInfeld DBI dynamics. Analytical features are reached through a firstorder formalism and a deformation procedure. The analysis ensures the linear stability of the obtained solutions, and the deformation method permits to detect new topological solutions given some systems of known solutions. The proposed models vary according to the parameters of the theory. However, in a certain parameter regime, their defect profiles are precisely obtained by standard theories. These are the models relatives. Besides that, we investigate the betaStarobinsky potential in the perspective of topological defects; and we have shown that it can support kinklike solutions, for both canonical and noncanonical kinetics. As a result, we propose two new kinds of generalizations on the betaStarobinsky model, by considering the DBI approach. Finally, we explore the main characteristics of such structures in these new scenarios.

1stOrder Dynamics on Nonlinear Agents for Resource Allocation over UniformlyConnected Networks ; A general nonlinear 1storder consensusbased solution for distributed constrained convex optimization is proposed with network resource allocation applications. The solution is used to optimize continuouslydifferentiable strictly convex cost functions over weaklyconnected undirected networks, while it is anytime feasible and models various nonlinearities to account for imperfections and constraints on the physical model of agents in terms of limited actuation capabilities, e.g., quantization and saturation. Due to such inherent nonlinearities, the existing linear solutions considering ideal agent models may not necessarily converge with guaranteed optimality and anytime feasibility. Some applications also impose specific nonlinearities, e.g., convergence in fixedfinitetime or signbased robust disturbancetolerant dynamics. Our proposed distributed protocol generalizes such nonlinear models. Putting convex set analysis together with nonsmooth Lyapunov analysis, we prove convergence, i regardless of the particular type of nonlinearity, and ii with weak networkconnectivity requirements uniformconnectivity.

Mathematical Modeling of Cell Collective Motion Triggered By SelfGenerated Gradients ; Selfgenerated gradients have atttracted a lot of attention in the recent biological literature. It is considered as a robust strategy for a group of cells to find its way during a long journey. This note is intended to discuss various scenarios for modeling traveling waves of cells that constantly deplete a chemical cue, and so create their own signaling gradient all along the way. We begin with one famous model by Keller and Segel for bacterial chemotaxis. We present the model and the construction of the traveling wave solutions. We also discuss the limitation of this approach, and review some subsequent work addressing stability issues. Next, we review two relevant extensions, which are supported by biological experiments. They both admit traveling wave solutions with an explicit value for the wave speed. We conclude by discussing some open problems and perspectives, and particularly a striking mechanism of speed determinacy occurring at the back of the wave. All the results presented in this note are illustrated by numerical simulations.

Deep Learning based Modelfree Robust Load Restoration to Enhance Bulk System Resilience with Wind Power Penetration ; This paper proposes a new deep learning DL based modelfree robust method for bulk system online load restoration with high penetration of wind power. Inspired by the iterative calculation of the twostage robust load restoration model, the deep neural network DNN and deep convolutional neural network CNN are respectively designed to find the worstcase system condition of a load pickup decision and evaluate the corresponding security. In order to find the optimal result within a limited number of checks, a load pickup checklist generation LPCG algorithm is developed to ensure the optimality. Then, the fast robust load restoration strategy acquisition is achieved based on the designed oneline strategy generation OSG algorithm. The proposed method finds the optimal result in a modelfree way, holds the robustness to handle uncertainties, and provides realtime computation. It can completely replace conventional robust optimization and supports online robust load restoration which better satisfies the changeable restoration process. The effectiveness of the proposed method is validated using the IEEE 30bus system and the IEEE 118bus system, showing high computational efficiency and considerable accuracy.

Economic Crises in a Model with Capital Scarcity and SelfReflexive Confidence ; In the General Theory, Keynes remarked that the economy's state depends on expectations, and that these expectations can be subject to sudden swings. In this work, we develop a multiple equilibria behavioural business cycle model that can account for demand or supply collapses due to abrupt drops in consumer confidence, which affect both consumption propensity and investment. We show that, depending on the model parameters, four qualitatively different outcomes can emerge, characterised by the frequency of capital scarcity andor demand crises. In the absence of policy measures, the duration of such crises can increase by orders of magnitude when parameters are varied, as a result of the paradox of thrift. Our model suggests policy recommendations that prevent the economy from getting trapped in extended stretches of low output, low investment and high unemployment.

Basic' Generalization Error Bounds for Least Squares Regression with Wellspecified Models ; This note examines the behavior of generalization capabilities as defined by outofsample mean squared error MSE of Linear Gaussian with a fixed design matrix and Linear Least Squares regression. Particularly, we consider a wellspecified model setting, i.e. we assume that there exists a true' combination of model parameters within the chosen model form. While the statistical properties of Least Squares regression have been extensively studied over the past few decades particularly with bf less restrictive problem statements compared to the present work this note targets bounds that are bf nonasymptotic and more quantitative compared to the literature. Further, the analytical formulae for distributions and bounds on the MSE are directly compared to numerical experiments. Derivations are presented in a selfcontained and pedagogical manner, in a way that a reader with a basic knowledge of probability and statistics can follow.

Modeling Adversarial Noise for Adversarial Training ; Deep neural networks have been demonstrated to be vulnerable to adversarial noise, promoting the development of defense against adversarial attacks. Motivated by the fact that adversarial noise contains wellgeneralizing features and that the relationship between adversarial data and natural data can help infer natural data and make reliable predictions, in this paper, we study to model adversarial noise by learning the transition relationship between adversarial labels i.e. the flipped labels used to generate adversarial data and natural labels i.e. the ground truth labels of the natural data. Specifically, we introduce an instancedependent transition matrix to relate adversarial labels and natural labels, which can be seamlessly embedded with the target model enabling us to model stronger adaptive adversarial noise. Empirical evaluations demonstrate that our method could effectively improve adversarial accuracy.

Mapping and Validating a Point Neuron Model on Intel's Neuromorphic Hardware Loihi ; Neuromorphic hardware is based on emulating the natural biological structure of the brain. Since its computational model is similar to standard neural models, it could serve as a computational acceleration for research projects in the field of neuroscience and artificial intelligence, including biomedical applications. However, in order to exploit this new generation of computer chips, rigorous simulation and consequent validation of brainbased experimental data is imperative. In this work, we investigate the potential of Intel's fifth generation neuromorphic chip Loihi', which is based on the novel idea of Spiking Neural Networks SNNs emulating the neurons in the brain. The work is implemented in context of simulating the Leaky Integrate and Fire LIF models based on the mouse primary visual cortex matched to a rich data set of anatomical, physiological and behavioral constraints. Simulations on the classical hardware serve as the validation platform for the neuromorphic implementation. We find that Loihi replicates classical simulations very efficiently and scales notably well in terms of both time and energy performance as the networks get larger.

Analysis and Methods to Mitigate Effects of Underreporting in Count Data ; Underreporting of count data poses a major roadblock for prediction and inference. In this paper, we focus on the Pogit model, which deconvolves the generating Poisson process from the censuring process controlling underreporting using a generalized linear modeling framework. We highlight the limitations of the Pogit model and address them by adding constraints to the estimation framework. We also develop uncertainty quantification techniques that are robust to model misspecification. Our approach is evaluated using synthetic data and applied to real healthcare datasets, where we treat inpatient data as reported' counts and use heldout total injuries to validate the results. The methods make it possible to separate the Poisson process from the underreporting process, given sufficient expert information. Codes to implement the approach are available via an open source Python package.

Cosmological Dynamics of CuscutaGalileon Gravity ; We study cosmological dynamics of the cuscutagalileon gravity with a potential term by using the dynamical system approach. This model is galileon generalization of the cuscuton gravity where we add a potential term to the theory in order to obtain the radiation and matter dominated eras. The exponential potential can provide the sequence of the thermal history of the Universe correctly, i.e. starting from radiation dominance, passing through matter dominant era, and then approaching de Sitter expansion stage. This model has no ghosts and the Laplacian instability for both scalar and tensor perturbations. We also discuss the observational constraints on the model parameters. It turns out that the model actually has three degrees of freedom unlike the original cuscuton theory.

Suboptimal nonlinear model predictive control with input moveblocking ; This paper deals with the integration of input moveblocking into the framework of suboptimal model predictive control. The blocked input parameterization is explicitly considered as a source of suboptimality. A straightforward integration approach is to hold back a manually generated stabilizing fallback solution in some buffer for the case that the optimizer does not find a better input moveblocked solution. An extended approach superimposes the manually generated stabilizing warmstart by the moveblocked control sequence and enables a stepwise improvement of the control performance. In addition, this contribution provides a detailed review of the literature on input moveblocked model predictive control and combines important results with the findings of suboptimal model predictive control. A numerical example supports the theoretical results and shows the effectiveness of the proposed approach.

MINIMAL Mining Models for Data Free Universal Adversarial Triggers ; It is well known that natural language models are vulnerable to adversarial attacks, which are mostly inputspecific in nature. Recently, it has been shown that there also exist inputagnostic attacks in NLP models, called universal adversarial triggers. However, existing methods to craft universal triggers are data intensive. They require large amounts of data samples to generate adversarial triggers, which are typically inaccessible by attackers. For instance, previous works take 3000 data samples per class for the SNLI dataset to generate adversarial triggers. In this paper, we present a novel datafree approach, MINIMAL, to mine inputagnostic adversarial triggers from models. Using the triggers produced with our datafree algorithm, we reduce the accuracy of Stanford Sentiment Treebank's positive class from 93.6 to 9.6. Similarly, for the Stanford Natural Language Inference SNLI, our singleword trigger reduces the accuracy of the entailment class from 90.95 to less than 0.6. Despite being completely datafree, we get equivalent accuracy drops as datadependent methods.

FlowVocoder A small Footprint Neural Vocoder based Normalizing flow for Speech Synthesis ; Recently, autoregressive neural vocoders have provided remarkable performance in generating highfidelity speech and have been able to produce synthetic speech in realtime. However, autoregressive neural vocoders such as WaveFlow are capable of modeling waveform signals from melspectrogram, its number of parameters is significant to deploy on edge devices. Though NanoFlow, which has a small number of parameters, is a stateoftheart autoregressive neural vocoder, the performance of NanoFlow is marginally lower than WaveFlow. Therefore, we propose a new type of autoregressive neural vocoder called FlowVocoder, which has a small memory footprint and is capable of generating highfidelity audio in realtime. Our proposed model improves the density estimation of flow blocks by utilizing a mixture of Cumulative Distribution Functions CDF for bipartite transformation. Hence, the proposed model is capable of modeling waveform signals, while its memory footprint is much smaller than WaveFlow. As shown in experiments, FlowVocoder achieves competitive results with baseline methods in terms of both subjective and objective evaluation, also, it is more suitable for realtime texttospeech applications.

Prose2Poem The Blessing of Transformers in Translating Prose to Persian Poetry ; Persian Poetry has consistently expressed its philosophy, wisdom, speech, and rationale on the basis of its couplets, making it an enigmatic language on its own to both native and nonnative speakers. Nevertheless, the notice able gap between Persian prose and poem has left the two pieces of literature mediumless. Having curated a parallel corpus of prose and their equivalent poems, we introduce a novel Neural Machine Translation NMT approach to translate prose to ancient Persian poetry using transformerbased Language Models in an extremely lowresource setting. More specifically, we trained a Transformer model from scratch to obtain initial translations and pretrained different variations of BERT to obtain final translations. To address the challenge of using masked language modelling under poeticness criteria, we heuristically joined the two models and generated valid poems in terms of automatic and human assessments. Final results demonstrate the eligibility and creativity of our novel heuristically aided approach among Literature professionals and nonprofessionals in generating novel Persian poems.