Datasets:

sswt

/

arxiv_sample_50K

like 0

Variational Disentanglement for Domain Generalization ; Domain generalization aims to learn an invariant model that can generalize well to the unseen target domain. In this paper, we propose to tackle the problem of domain generalization by delivering an effective framework named Variational Disentanglement Network VDN, which is capable of disentangling the domainspecific features and taskspecific features, where the taskspecific features are expected to be better generalized to unseen but related test data. We further show the rationale of our proposed method by proving that our proposed framework is equivalent to minimize the evidence upper bound of the divergence between the distribution of taskspecific features and its invariant ground truth derived from variational inference. We conduct extensive experiments to verify our method on three benchmarks, and both quantitative and qualitative results illustrate the effectiveness of our method.

Generalized XGBoost Method ; The XGBoost method has many advantages and is especially suitable for statistical analysis of big data, but its loss function is limited to convex functions. In many specific applications, a nonconvex loss function would be preferable. In this paper, I propose a generalized XGBoost method, which requires weaker loss function constraint and involves more general loss functions, including convex loss functions and some nonconvex loss functions. Furthermore, this generalized XGBoost method is extended to multivariate loss function to form a more generalized XGBoost method. This method is a multiobjective parameter regularized tree boosting method, which can model multiple parameters in most of the frequentlyused parametric probability distributions to be fitted by predictor variables. Meanwhile, the related algorithms and some examples in nonlife insurance pricing are given.

Horndeski Proca stars with vector hair ; We study Proca stars in a vectortensor gravity model inspired by Horndeski's generalized EinsteinMaxwell field equations, supplemented with a mass term for the vector field. We discuss the effects of the nonminimal coupling term on the properties of the resulting Proca stars. and show that they can form vector hair. We show that the sign of the coupling constant is crucial in determining the generic properties of these generalized Proca star solutions, as soon as the magnitude of the coupling constant is sufficiently large to allow for significant deviations from the standard Proca star case. For negative coupling constant we observe a new type of limiting behavior for the generalized Proca stars, where the spacetime splits into an interior region with matter fields and an exterior Schwarzschild region.

Learning to Selectively Learn for Weaklysupervised Paraphrase Generation ; Paraphrase generation is a longstanding NLP task that has diverse applications for downstream NLP tasks. However, the effectiveness of existing efforts predominantly relies on large amounts of golden labeled data. Though unsupervised endeavors have been proposed to address this issue, they may fail to generate meaningful paraphrases due to the lack of supervision signals. In this work, we go beyond the existing paradigms and propose a novel approach to generate highquality paraphrases with weak supervision data. Specifically, we tackle the weaklysupervised paraphrase generation problem by 1 obtaining abundant weaklylabeled parallel sentences via retrievalbased pseudo paraphrase expansion; and 2 developing a metalearning framework to progressively select valuable samples for finetuning a pretrained language model, i.e., BART, on the sentential paraphrasing task. We demonstrate that our approach achieves significant improvements over existing unsupervised approaches, and is even comparable in performance with supervised stateofthearts.

Sparse Data Generation for ParticleBased Simulation of Hadronic Jets in the LHC ; We develop a generative neural network for the generation of sparse data in particle physics using a permutationinvariant and physicsinformed loss function. The input dataset used in this study consists of the particle constituents of hadronic jets due to its sparsity and the possibility of evaluating the network's ability to accurately describe the particles and jets properties. A variational autoencoder composed of convolutional layers in the encoder and decoder is used as the generator. The loss function consists of a reconstruction error term and the KullbackLeibler divergence between the output of the encoder and the latent vector variables. The permutationinvariant loss on the particles' properties is combined with two meansquared error terms that measure the difference between input and output jets mass and transverse momentum, which improves the network's generation capability as it imposes physics constraints, allowing the model to learn the kinematics of the jets.

Accurate relativistic observables from postprocessing light cone catalogues ; We introduce and study a new scheme to construct relativistic observables from postprocessing light cone data. This construction is based on a novel approach, LCMetric, which takes general light cone or snapshot output generated by arbitrary Nbody simulations or emulations and solves the linearized Einstein equations to determine the spacetime metric on the light cone. We find that this scheme is able to determine the metric to high precision, and subsequently generate accurate mock cosmological observations sensitive to effects such as postBorn lensing and nonlinear ISW contributions. By comparing to conventional methods in quantifying those general relativistic effects, we show that this scheme is able to accurately construct the lensing convergence signal. We also find the accuracy of this method in quantifying the ISW effects in the highly nonlinear regime outperforms conventional methods by an order of magnitude. This scheme opens a new path for exploring and modeling higherorder and nonlinear general relativistic contributions to cosmological observables, including mock observations of gravitational lensing and the moving lens and ReesSciama effects.

Exploring Conditional Text Generation for AspectBased Sentiment Analysis ; Aspectbased sentiment analysis ABSA is an NLP task that entails processing usergenerated reviews to determine i the target being evaluated, ii the aspect category to which it belongs, and iii the sentiment expressed towards the target and aspect pair. In this article, we propose transforming ABSA into an abstract summarylike conditional text generation task that uses targets, aspects, and polarities to generate auxiliary statements. To demonstrate the efficacy of our task formulation and a proposed system, we finetune a pretrained model for conditional text generation tasks to get new stateoftheart results on a few restaurant domains and urban neighborhoods domain benchmark datasets.

Distinguishing rule and exemplarbased generalization in learning systems ; Machine learning systems often do not share the same inductive biases as humans and, as a result, extrapolate or generalize in ways that are inconsistent with our expectations. The tradeoff between exemplar and rulebased generalization has been studied extensively in cognitive psychology; in this work, we present a protocol inspired by these experimental approaches to probe the inductive biases that control this tradeoff in categorylearning systems. We isolate two such inductive biases featurelevel bias differences in which features are more readily learned and exemplar or rule bias differences in how these learned features are used for generalization. We find that standard neural network models are featurebiased and exemplarbased, and discuss the implications of these findings for machine learning research on systematic generalization, fairness, and data augmentation.

Omnidata A Scalable Pipeline for Making MultiTask MidLevel Vision Datasets from 3D Scans ; This paper introduces a pipeline to parametrically sample and render multitask vision datasets from comprehensive 3D scans from the real world. Changing the sampling parameters allows one to steer the generated datasets to emphasize specific information. In addition to enabling interesting lines of research, we show the tooling and generated data suffice to train robust vision models. Common architectures trained on a generated starter dataset reached stateoftheart performance on multiple common vision tasks and benchmarks, despite having seen no benchmark or nonpipeline data. The depth estimation network outperforms MiDaS and the surface normal estimation network is the first to achieve humanlevel performance for inthewild surface normal estimation at least according to one metric on the OASIS benchmark. The Dockerized pipeline with CLI, the mostly python code, PyTorch dataloaders for the generated data, the generated starter dataset, download scripts and other utilities are available through our project website, httpsomnidata.vision.

A Candidate for a Supergravity Anomaly ; Using elementary BRS cohomology theory, this paper describes a supergravity anomaly analogous to, but very different from, the well known gauge and gravitational anomalies. It closely resembles the known gauge anomalies, but it results from a triangle diagram with two gravitinos and a gauge vector boson, rather than three gauge vector bosons, or two gravitons and a vector boson. A model that is likely to generate this supergravity anomaly is described. The coefficient of this anomaly, in perturbation theory, in a theory with unbroken supersymmetry, appears to be zero, because no relevant diagrams are linearly divergent. However, when, and only when, the theory has spontaneously broken supergravity, there are counterterms in the action which contribute to linearly divergent diagrams that can generate the anomaly. From the relevant Feynman diagrams in the theory, the general form of the anomaly can be conjectured. It is proportional to the VEV left Da right of the auxiliary field for the vector boson. So removing the anomaly generates a requirement that the effective spontaneous breaking of supergravity needs to be of the purely chiral type with left Fi right not 0 and with left Da right0.

Molecular Graph Generation via Geometric Scattering ; Graph neural networks GNNs have been used extensively for addressing problems in drug design and discovery. Both ligand and target molecules are represented as graphs with node and edge features encoding information about atomic elements and bonds respectively. Although existing deep learning models perform remarkably well at predicting physicochemical properties and binding affinities, the generation of new molecules with optimized properties remains challenging. Inherently, most GNNs perform poorly in wholegraph representation due to the limitations of the messagepassing paradigm. Furthermore, stepbystep graph generation frameworks that use reinforcement learning or other sequential processing can be slow and result in a high proportion of invalid molecules with substantial postprocessing needed in order to satisfy the principles of stoichiometry. To address these issues, we propose a representationfirst approach to molecular graph generation. We guide the latent representation of an autoencoder by capturing graph structure information with the geometric scattering transform and apply penalties that structure the representation also by molecular properties. We show that this highly structured latent space can be directly used for molecular graph generation by the use of a GAN. We demonstrate that our architecture learns meaningful representations of drug datasets and provides a platform for goaldirected drug synthesis.

Amortized Tree Generation for Bottomup Synthesis Planning and Synthesizable Molecular Design ; Molecular design and synthesis planning are two critical steps in the process of molecular discovery that we propose to formulate as a single shared task of conditional synthetic pathway generation. We report an amortized approach to generate synthetic pathways as a Markov decision process conditioned on a target molecular embedding. This approach allows us to conduct synthesis planning in a bottomup manner and design synthesizable molecules by decoding from optimized conditional codes, demonstrating the potential to solve both problems of design and synthesis simultaneously. The approach leverages neural networks to probabilistically model the synthetic trees, one reaction step at a time, according to reactivity rules encoded in a discrete action space of reaction templates. We train these networks on hundreds of thousands of artificial pathways generated from a pool of purchasable compounds and a list of expertcurated templates. We validate our method with a the recovery of molecules using conditional generation, b the identification of synthesizable structural analogs, and c the optimization of molecular structures given oracle functions relevant to drug discovery.

StyleNeRF A Stylebased 3DAware Generator for Highresolution Image Synthesis ; We propose StyleNeRF, a 3Daware generative model for photorealistic highresolution image synthesis with high multiview consistency, which can be trained on unstructured 2D images. Existing approaches either cannot synthesize highresolution images with fine details or yield noticeable 3Dinconsistent artifacts. In addition, many of them lack control over style attributes and explicit 3D camera poses. StyleNeRF integrates the neural radiance field NeRF into a stylebased generator to tackle the aforementioned challenges, i.e., improving rendering efficiency and 3D consistency for highresolution image generation. We perform volume rendering only to produce a lowresolution feature map and progressively apply upsampling in 2D to address the first issue. To mitigate the inconsistencies caused by 2D upsampling, we propose multiple designs, including a better upsampler and a new regularization loss. With these designs, StyleNeRF can synthesize highresolution images at interactive rates while preserving 3D consistency at high quality. StyleNeRF also enables control of camera poses and different levels of styles, which can generalize to unseen views. It also supports challenging tasks, including zoomin andout, style mixing, inversion, and semantic editing.

Generating Symbolic Reasoning Problems with Transformer GANs ; We study the capabilities of GANs and Wasserstein GANs equipped with Transformer encoders to generate sensible and challenging training data for symbolic reasoning domains. We conduct experiments on two problem domains where Transformers have been successfully applied recently symbolic mathematics and temporal specifications in verification. Even without autoregression, our GAN models produce syntactically correct instances. We show that the generated data can be used as a substitute for real training data when training a classifier, and, especially, that training data can be generated from a dataset that is too small to be trained on directly. Using a GAN setting also allows us to alter the target distribution We show that by adding a classifier uncertainty part to the generator objective, we obtain a dataset that is even harder to solve for a temporal logic classifier than our original dataset.

Twodimensional mesh generator in generalized coordinates implemented in Python ; Through mathematical models, it is possible to turn a problem of the physical domain into the computational domain. In this context, the paper presents a twodimensional mesh generator in generalized coordinates, which uses the Parametric Linear Spline method and partial differential equations. The generator is automated and able to treat real complex domains. The code was implemented in Python, applying the Numpy and Matplotlib libraries to matrix manipulations and graphical plots, respectively. Applications are made for monoblock meshes twodimensional shape of a bottle and multiblock meshes geometry of Igap'o I lake, Londrina, Paran'a, Brazil.

Characterization of Time Delay Interferometry combinations for the LISA instrument noise ; Time delay interferometry TDI is a postprocessing technique used in the Laser Interferometer Space Antenna LISA to reduce laser frequency noise by building an equalarm interferometer via combining timeshifted raw phase measurements. The set of socalled 2nd generation TDI variables which sufficiently suppress laser frequency noise considering realistic LISA orbital dynamics has recently been expanded by a large number of additional solutions. In this paper, we characterize these new TDI channels by relating them to the wellknown 1st generation variables alpha, beta, gamma, and zeta. We compute explicitly how each 2nd generation variable can be approximated as a linear combination of these four 1st generation variables, and show numerically that these approximations are accurate enough to model the noises not suppressed by TDI. We use these results to discuss how the newly found channels might be advantageous to use for the LISA data analysis. In addition, we demonstrate that newly found variants of the variable zeta significantly outperform the ones previously known from the literature.

A Theoretical Analysis on Independencedriven Importance Weighting for Covariateshift Generalization ; Covariateshift generalization, a typical case in outofdistribution OOD generalization, requires a good performance on the unknown test distribution, which varies from the accessible training distribution in the form of covariate shift. Recently, independencedriven importance weighting algorithms in stable learning literature have shown empirical effectiveness to deal with covariateshift generalization on several learning models, including regression algorithms and deep neural networks, while their theoretical analyses are missing. In this paper, we theoretically prove the effectiveness of such algorithms by explaining them as feature selection processes. We first specify a set of variables, named minimal stable variable set, that is the minimal and optimal set of variables to deal with covariateshift generalization for common loss functions, such as the mean squared loss and binary crossentropy loss. Afterward, we prove that under ideal conditions, independencedriven importance weighting algorithms could identify the variables in this set. Analysis of asymptotic properties is also provided. These theories are further validated in several synthetic experiments.

Deceive D Adaptive Pseudo Augmentation for GAN Training with Limited Data ; Generative adversarial networks GANs typically require ample data for training in order to synthesize highfidelity images. Recent studies have shown that training GANs with limited data remains formidable due to discriminator overfitting, the underlying cause that impedes the generator's convergence. This paper introduces a novel strategy called Adaptive Pseudo Augmentation APA to encourage healthy competition between the generator and the discriminator. As an alternative method to existing approaches that rely on standard data augmentations or model regularization, APA alleviates overfitting by employing the generator itself to augment the real data distribution with generated images, which deceives the discriminator adaptively. Extensive experiments demonstrate the effectiveness of APA in improving synthesis quality in the lowdata regime. We provide a theoretical analysis to examine the convergence and rationality of our new training strategy. APA is simple and effective. It can be added seamlessly to powerful contemporary GANs, such as StyleGAN2, with negligible computational cost.

HydraGAN A Multihead, Multiobjective Approach to Synthetic Data Generation ; Synthetic data generation overcomes limitations of realworld machine learning. Traditional methods are valuable for augmenting costly datasets but only optimize one criterion realism. In this paper, we tackle the problem of generating synthetic data that optimize multiple criteria. This goal is necessary when real data are replaced by synthetic for privacy preservation. We introduce HydraGAN, a new approach to synthetic data generation that introduces multiple generator and discriminator agents into the system. The multiagent GAN optimizes the goal of privacypreservation as well as data realism. To facilitate multiagent training, we adapt gametheoretic principles to offer equilibrium guarantees. We observe that HydraGAN outperforms baseline methods for three datasets for multiple criteria of maximizing data realism, maximizing model accuracy, and minimizing reidentification risk.

AMuzeNet Music Generation by Composing the Harmony based on the Generated Melody ; We present a method for the generation of Midi files of piano music. The method models the right and left hands using two networks, where the left hand is conditioned on the right hand. This way, the melody is generated before the harmony. The Midi is represented in a way that is invariant to the musical scale, and the melody is represented, for the purpose of conditioning the harmony, by the content of each bar, viewed as a chord. Finally, notes are added randomly, based on this chord representation, in order to enrich the generated audio. Our experiments show a significant improvement over the state of the art for training on such datasets, and demonstrate the contribution of each of the novel components.

SemanticAware Generation for SelfSupervised Visual Representation Learning ; In this paper, we propose a selfsupervised visual representation learning approach which involves both generative and discriminative proxies, where we focus on the former part by requiring the target network to recover the original image based on the midlevel features. Different from prior work that mostly focuses on pixellevel similarity between the original and generated images, we advocate for Semanticaware Generation SaGe to facilitate richer semantics rather than details to be preserved in the generated image. The core idea of implementing SaGe is to use an evaluator, a deep network that is pretrained without labels, for extracting semanticaware features. SaGe complements the target network with viewspecific features and thus alleviates the semantic degradation brought by intensive data augmentations. We execute SaGe on ImageNet1K and evaluate the pretrained models on five downstream tasks including nearest neighbor test, linear classification, and finescaled image recognition, demonstrating its ability to learn stronger visual representations.

Towards Principled Disentanglement for Domain Generalization ; A fundamental challenge for machine learning models is generalizing to outofdistribution OOD data, in part due to spurious correlations. To tackle this challenge, we first formalize the OOD generalization problem as constrained optimization, called Disentanglementconstrained Domain Generalization DDG. We relax this nontrivial constrained optimization problem to a tractable form with finitedimensional parameterization and empirical approximation. Then a theoretical analysis of the extent to which the above transformations deviates from the original problem is provided. Based on the transformation, we propose a primaldual algorithm for joint representation disentanglement and domain generalization. In contrast to traditional approaches based on domain adversarial training and domain labels, DDG jointly learns semantic and variation encoders for disentanglement, enabling flexible manipulation and augmentation on training data. DDG aims to learn intrinsic representations of semantic concepts that are invariant to nuisance factors and generalizable across domains. Comprehensive experiments on popular benchmarks show that DDG can achieve competitive OOD performance and uncover interpretable salient structures within data.

Just Least Squares Binary Compressive Sampling with Low Generative Intrinsic Dimension ; In this paper, we consider recovering n dimensional signals from m binary measurements corrupted by noises and sign flips under the assumption that the target signals have low generative intrinsic dimension, i.e., the target signals can be approximately generated via an LLipschitz generator G mathbbRkrightarrowmathbbRn, kll n. Although the binary measurements model is highly nonlinear, we propose a least square decoder and prove that, up to a constant c, with high probability, the least square decoder achieves a sharp estimation error mathcalO sqrtfracklog Lnm as long as mgeq mathcalO klog Ln. Extensive numerical simulations and comparisons with stateoftheart methods demonstrated the least square decoder is robust to noise and sign flips, as indicated by our theory. By constructing a ReLU network with properly chosen depth and width, we verify the approximately deep generative prior, which is of independent interest.

WideSense Stationarity in Generalized Graph Signal Processing ; We consider statistical graph signal processing GSP in a generalized framework where each vertex of a graph is associated with an element from a Hilbert space. This general model encompasses various signals such as the traditional scalarvalued graph signal, multichannel graph signal, and discrete and continuoustime graph signals, allowing us to build a unified theory of graph random processes. We introduce the notion of joint widesense stationarity in this generalized GSP framework, which allows us to characterize a graph random process as a combination of uncorrelated oscillation modes across both the vertex and Hilbert space domains. We elucidate the relationship between the notions of widesense stationarity in different domains, and derive the Wiener filters for denoising and signal completion under this framework. Numerical experiments on both real and synthetic datasets demonstrate the utility of our generalized approach in achieving better estimation performance compared to traditional GSP or the timevertex framework.

Adaptive Feature Interpolation for LowShot Image Generation ; Training of generative models especially Generative Adversarial Networks can easily diverge in lowdata setting. To mitigate this issue, we propose a novel implicit data augmentation approach which facilitates stable training and synthesize highquality samples without need of label information. Specifically, we view the discriminator as a metric embedding of the real data manifold, which offers proper distances between real data points. We then utilize information in the feature space to develop a fully unsupervised and datadriven augmentation method. Experiments on fewshot generation tasks show the proposed method significantly improve results from strong baselines with hundreds of training samples.

The Price Impact of Generalized Order Flow Imbalance ; Order flow imbalance can explain shortterm changes in stock price. This paper considers the change of nonminimum quotation units in real transactions, and proposes a generalized order flow imbalance construction method to improve Order Flow Imbalance OFI and Stationarized Order Flow Imbalance logOFI. Based on the highfrequency order book snapshot data, we conducted an empirical analysis of the CSI 500 constituent stocks. In order to facilitate the presentation, we selected 10 stocks for comparison. The two indicators after the improvement of the generalized order flow imbalance construction method both show a better ability to explain changes in stock prices. Especially Generalized Stationarized Order Flow Imbalance logGOFI, using a linear regression model, on the time scales of 30 seconds, 1 minute, and 5 minutes, the average Rsquared out of sample compared with Order Flow Imbalance OFI 32.89, 38.13 and 42.57, respectively increased to 83.57, 85.37 and 86.01. In addition, we found that the interpretability of Generalized Stationarized Order Flow Imbalance logGOFI showed stronger stability on all three time scales.

ViewCLR Learning Selfsupervised Video Representation for Unseen Viewpoints ; Learning selfsupervised video representation predominantly focuses on discriminating instances generated from simple data augmentation schemes. However, the learned representation often fails to generalize over unseen camera viewpoints. To this end, we propose ViewCLR, that learns selfsupervised video representation invariant to camera viewpoint changes. We introduce a viewgenerator that can be considered as a learnable augmentation for any selfsupervised pretext tasks, to generate latent viewpoint representation of a video. ViewCLR maximizes the similarities between the latent viewpoint representation with its representation from the original viewpoint, enabling the learned video encoder to generalize over unseen camera viewpoints. Experiments on crossview benchmark datasets including NTU RGBD dataset show that ViewCLR stands as a stateoftheart viewpoint invariant selfsupervised method.

Nonadiabatic quantum dynamics of tribovoltaic effects at sliding metalsemiconductor interfaces ; Recent experiments observe electric current generation at a sliding metalsemiconductor interfaces. Here, we present a detailed theoretical study on how electric voltage is generated at such a sliding interface. Our study is based on a twoband AndersonHolstein model, and we solve the coupled electronphonon dynamics using a surface hopping method. We show that the high local temperature induced by mechanic motion at the interfaces could lead to electronhole pair generation through electronphonon couplings. We quantify the efficiency of electronhole generation as well as electric voltage as a function of local temperatures and semiconductor bandgaps. We find that increasing the local temperatures can lead to higher electronhole generations and electric voltage. Furthermore, we find that there is a turnover for the electric voltage as a function of the bandgap. Such an observation is in agreement with the experimental results. Our study offers a theoretical framework to understand tribovoltaic effects from a quantum mechanical point of view, and our approach can be used to quantitively simulate realistic sliding metalsemiconductor junctions.

A Shared Representation for Photorealistic Driving Simulators ; A powerful simulator highly decreases the need for realworld tests when training and evaluating autonomous vehicles. Datadriven simulators flourished with the recent advancement of conditional Generative Adversarial Networks cGANs, providing highfidelity images. The main challenge is synthesizing photorealistic images while following given constraints. In this work, we propose to improve the quality of generated images by rethinking the discriminator architecture. The focus is on the class of problems where images are generated given semantic inputs, such as scene segmentation maps or human body poses. We build on successful cGAN models to propose a new semanticallyaware discriminator that better guides the generator. We aim to learn a shared latent representation that encodes enough information to jointly do semantic segmentation, content reconstruction, along with a coarsetofine grained adversarial reasoning. The achieved improvements are generic and simple enough to be applied to any architecture of conditional image synthesis. We demonstrate the strength of our method on the scene, building, and human synthesis tasks across three different datasets. The code is available at httpsgithub.comvitaepflSemDisc.

GM Score Incorporating interclass and intraclass generator diversity, discriminability of disentangled representation, and sample fidelity for evaluating GANs ; While generative adversarial networks GAN are popular for their higher sample quality as opposed to other generative models like the variational autoencoders VAE and Boltzmann machines, they suffer from the same difficulty of the evaluation of generated samples. Various aspects must be kept in mind, such as the quality of generated samples, the diversity of classes within a class and among classes, the use of disentangled latent spaces, agreement of said evaluation metric with human perception, etc. In this paper, we propose a new score, namely, GM Score, which takes into various factors such as sample quality, disentangled representation, intraclass and interclass diversity, and other metrics such as precision, recall, and F1 score are employed for discriminability of latent space of deep belief network DBN and restricted Boltzmann machine RBM. The evaluation is done for different GANs GAN, DCGAN, BiGAN, CGAN, CoupledGAN, LSGAN, SGAN, WGAN, and WGAN Improved trained on the benchmark MNIST dataset.

DFANeRF Personalized Talking Head Generation via Disentangled Face Attributes Neural Rendering ; While recent advances in deep neural networks have made it possible to render highquality images, generating photorealistic and personalized talking head remains challenging. With given audio, the key to tackling this task is synchronizing lip movement and simultaneously generating personalized attributes like head movement and eye blink. In this work, we observe that the input audio is highly correlated to lip motion while less correlated to other personalized attributes e.g., head movements. Inspired by this, we propose a novel framework based on neural radiance field to pursue highfidelity and personalized talking head generation. Specifically, neural radiance field takes lip movements features and personalized attributes as two disentangled conditions, where lip movements are directly predicted from the audio inputs to achieve lipsynchronized generation. In the meanwhile, personalized attributes are sampled from a probabilistic model, where we design a Transformerbased variational autoencoder sampled from Gaussian Process to learn plausible and naturallooking head pose and eye blink. Experiments on several benchmarks demonstrate that our method achieves significantly better results than stateoftheart methods.

Conditional Generation of Medical Time Series for Extrapolation to Underrepresented Populations ; The widespread adoption of electronic health records EHRs and subsequent increased availability of longitudinal healthcare data has led to significant advances in our understanding of health and disease with direct and immediate impact on the development of new diagnostics and therapeutic treatment options. However, access to EHRs is often restricted due to their perceived sensitive nature and associated legal concerns, and the cohorts therein typically are those seen at a specific hospital or network of hospitals and therefore not representative of the wider population of patients. Here, we present HealthGen, a new approach for the conditional generation of synthetic EHRs that maintains an accurate representation of real patient characteristics, temporal information and missingness patterns. We demonstrate experimentally that HealthGen generates synthetic cohorts that are significantly more faithful to real patient EHRs than the current stateoftheart, and that augmenting real data sets with conditionally generated cohorts of underrepresented subpopulations of patients can significantly enhance the generalisability of models derived from these data sets to different patient populations. Synthetic conditionally generated EHRs could help increase the accessibility of longitudinal healthcare data sets and improve the generalisability of inferences made from these data sets to underrepresented populations.

Hexagonal and trigonal quasiperiodic tilings ; Exploring nonminimalrank quasicrystals, which have symmetries that can be found in both periodic and aperiodic crystals, often provides new insight into the physical nature of aperiodic longrange order in models that are easier to treat. Motivated by the prevalence of experimental systems exhibiting aperiodic longrange order with hexagonal and trigonal symmetry, we introduce a generic twoparameter family of 2dimensional quasiperiodic tilings with such symmetries. We focus on the special case of trigonal and hexagonal Fibonacci, or goldenmean, tilings, analogous to the well studied square Fibonacci tiling. We first generate the tilings using a generalized version of de Bruijn's dual grid method. We then discuss their interpretation in terms of projections of a hypercubic lattice from six dimensional superspace. We conclude by concentrating on two of the hexagonal members of the family, and examining a few of their properties more closely, while providing a set of substitution rules for their generation.

ItoWave Ito Stochastic Differential Equation Is All You Need For Wave Generation ; In this paper, we propose a vocoder based on a pair of forward and reversetime linear stochastic differential equations SDE. The solutions of this SDE pair are two stochastic processes, one of which turns the distribution of wave, that we want to generate, into a simple and tractable distribution. The other is the generation procedure that turns this tractable simple signal into the target wave. The model is called ItoWave. ItoWave use the Wiener process as a driver to gradually subtract the excess signal from the noise signal to generate realistic corresponding meaningful audio respectively, under the conditional inputs of original mel spectrogram. The results of the experiment show that the mean opinion scores MOS of ItoWave can exceed the current stateoftheart SOTA methods, and reached 4.35pm0.115. The generated audio samples are available online.

A Dynamical Systems Framework for Generating the Riemann Zeta Function and Dirichlet Lfunctions ; We first construct a dynamical systems model which in its steadystate serves as an analytic continuation of the completed Riemann zeta function over the entire critical strip. The resulting mathematical construct involves a linear interpolation of two symmetric generator functions which can be used to infer the global properties of the nontrivial zeros of the Riemann zeta function using concentration bounds. The proposed dynamical systems framework thus provides an alternative method for investigating the celebrated Riemann Hypothesis which is shown in this paper to be almost surely true. We also show that the framework is general enough to study the nontrivial zeros of the Dirichlet Lfunctions and in this paper we show that under specific conditions, the generalized Riemann Hypothesis is also almost surely true.

Phasematched highorder harmonic generation in preionized noble gases ; One of the main difficulties to efficiently generating highorder harmonics in long neutralgas targets is to reach the phasematching conditions. One issue is that the medium cannot be sufficiently ionized by the driving laser due to plasma defocusing. We propose a method to improve the phasematching by preionizing the gas using a weak capillary discharge. We have demonstrated this mechanism, for the first time, in absorptionlimited XUV generation by an 800 nm femtosecond laser in argon and krypton. The phasemismatch control ability of our method is confirmed by an analytical model and numerical simulation of the complete generation process. Our method allows increasing the efficiency of the harmonic generation significantly, paving the way towards photonhungry applications of these shortwavelength compact sources.

MusIAC An extensible generative framework for Music Infilling Applications with multilevel Control ; We present a novel music generation framework for music infilling, with a user friendly interface. Infilling refers to the task of generating musical sections given the surrounding multitrack music. The proposed transformerbased framework is extensible for new control tokens as the added music control tokens such as tonal tension per bar and track polyphony level in this work. We explore the effects of including several musically meaningful control tokens, and evaluate the results using objective metrics related to pitch and rhythm. Our results demonstrate that adding additional control tokens helps to generate music with stronger stylistic similarities to the original music. It also provides the user with more control to change properties like the music texture and tonal tension in each bar compared to previous research which only provided control for track density. We present the model in a Google Colab notebook to enable interactive generation.

Multilevel Latent Space Structuring for Generative Control ; Truncation is widely used in generative models for improving the quality of the generated samples, at the expense of reducing their diversity. We propose to leverage the StyleGAN generative architecture to devise a new truncation technique, based on a decomposition of the latent space into clusters, enabling customized truncation to be performed at multiple semantic levels. We do so by learning to regenerate Wspace, the extended intermediate latent space of StyleGAN, using a learnable mixture of Gaussians, while simultaneously training a classifier to identify, for each latent vector, the cluster that it belongs to. The resulting truncation scheme is more faithful to the original untruncated samples and allows a better tradeoff between quality and diversity. We compare our method to other truncation approaches for StyleGAN, both qualitatively and quantitatively.

Generalized Matching Condition for Unity Efficiency Quantum Transduction ; Coherently converting quantum states between distinct elements via quantum transducers remains a crucial yet challenging task in quantum science. Especially in demand is quantum transduction between optical frequencies, which are ideal for lowloss transmission across long distances, and microwave frequencies, which admit highfidelity quantum operations. We present a generic formalism for Nstage quantum transduction that covers various leading microwavetooptical, microwavetomicrowave, and opticaltooptical linear conversion approaches. We then identify effective circuit models and the resulting generalized matching conditions for achieving maximum conversion efficiency. The generalized matching condition requires resistance matching as well as frequency matching beyond the usual resonant assumption, with simple impedancematched transmission interpretation. Our formalism provides a generic toolbox for determining experimental parameters to realize efficient quantum transduction, and suggests new regimes of nonresonant conversions that might outperform allresonant ones.

GANgenerated Faces Detection A Survey and New Perspectives ; Generative Adversarial Networks GAN have led to the generation of very realistic face images, which have been used in fake social media accounts and other disinformation matters that can generate profound impacts. Therefore, the corresponding GANface detection techniques are under active development that can examine and expose such fake faces. In this work, we aim to provide a comprehensive review of recent progress in GANface detection. We focus on methods that can detect face images that are generated or synthesized from GAN models. We classify the existing detection works into four categories 1 deep learningbased, 2 physicalbased, 3 physiologicalbased methods, and 4 evaluation and comparison against human visual performance. For each category, we summarize the key ideas and connect them with method implementations. We also discuss open problems and suggest future research directions.

Combining Varied Learners for Binary Classification using Stacked Generalization ; The Machine Learning has various learning algorithms that are better in some or the other aspect when compared with each other but a common error that all algorithms will suffer from is training data with very high dimensional feature set. This usually ends up algorithms into generalization error that deplete the performance. This can be solved using an Ensemble Learning method known as Stacking commonly termed as Stacked Generalization. In this paper we perform binary classification using Stacked Generalization on high dimensional Polycystic Ovary Syndrome dataset and prove the point that model becomes generalized and metrics improve significantly. The various metrics are given in this paper that also point out a subtle transgression found with Receiver Operating Characteristic Curve that was proved to be incorrect.

Diversity in deep generative models and generative AI ; The machine learning generative algorithms such as Generative Adversarial Networks GAN and Variational AutoEncoders VAE show impressive results when constructing objects similar to those in a training ensemble. However, the generation of new objects builds mainly on the understanding of the hidden structure of the training dataset followed by a sampling from a multidimensional normal variable. In particular each sample is independent from the others and can repeatedly propose same type of objects. To cure this drawback we introduce a kernelbased measure quantization method that can produce new objects from a given target measure by approximating it as a whole and even staying away from elements already drawn from that distribution. This ensures a better diversity of the produced objects. The method is tested on classic machine learning benchmarks.

DoCoGen Domain Counterfactual Generation for Low Resource Domain Adaptation ; Natural language processing NLP algorithms have become very successful, but they still struggle when applied to outofdistribution examples. In this paper we propose a controllable generation approach in order to deal with this domain adaptation DA challenge. Given an input text example, our DoCoGen algorithm generates a domaincounterfactual textual example Dcon that is similar to the original in all aspects, including the task label, but its domain is changed to a desired one. Importantly, DoCoGen is trained using only unlabeled examples from multiple domains no NLP task labels or parallel pairs of textual examples and their domaincounterfactuals are required. We show that DoCoGen can generate coherent counterfactuals consisting of multiple sentences. We use the Dcons generated by DoCoGen to augment a sentiment classifier and a multilabel intent classifier in 20 and 78 DA setups, respectively, where sourcedomain labeled data is scarce. Our model outperforms strong baselines and improves the accuracy of a stateoftheart unsupervised DA algorithm.

Pix2NeRF Unsupervised Conditional GAN for Single Image to Neural Radiance Fields Translation ; We propose a pipeline to generate Neural Radiance FieldsNeRF of an object or a scene of a specific class, conditioned on a single input image. This is a challenging task, as training NeRF requires multiple views of the same scene, coupled with corresponding poses, which are hard to obtain. Our method is based on piGAN, a generative model for unconditional 3Daware image synthesis, which maps random latent codes to radiance fields of a class of objects. We jointly optimize 1 the piGAN objective to utilize its highfidelity 3Daware generation and 2 a carefully designed reconstruction objective. The latter includes an encoder coupled with piGAN generator to form an autoencoder. Unlike previous fewshot NeRF approaches, our pipeline is unsupervised, capable of being trained with independent images without 3D, multiview, or pose supervision. Applications of our pipeline include 3d avatar generation, objectcentric novel view synthesis with a single input image, and 3daware superresolution, to name a few.

Perturbations in NonFlat Cosmology for fT gravity ; The study of cosmological perturbation theory in fT gravity is a topic of great interest in teleparallel gravity since this is one of the simplest generalizations of the theory that modifies the teleparallel equivalent of general relativity. In this work, we explore the possibility of a nonflat FLRW background solution and perform perturbations for positively as well as negatively curved spatial geometries, together with a comparison to the flat case. We determine the generalized behaviour of the perturbative modes for this nonflat FLRW setting for arbitrary fT models, when the most general homogeneous and isotropic background tetrads are used. We also identify propagating modes in this setup, and relate this with the case of a flat cosmology.

Fractional Calderon problems and Poincare inequalities on unbounded domains ; We generalize many recent uniqueness results on the fractional Calder'on problem to cover the cases of all domains with nonempty exterior. The highlight of our work is the characterization of uniqueness and nonuniqueness of partial data inverse problems for the fractional conductivity equation on domains that are bounded in one direction for conductivities supported in the whole Euclidean space and decaying to a constant background conductivity at infinity. We generalize the uniqueness proof for the fractional Calder'on problem by Ghosh, Salo and Uhlmann to a general abstract setting in order to use the full strength of their argument. This allows us to observe that there are also uniqueness results for many inverse problems for higher order local perturbations of a lower order fractional Laplacian. We give concrete example models to illustrate these curious situations and prove Poincar'e inequalities for the fractional Laplacians of any order on domains that are bounded in one direction. We establish Runge approximation results in these general settings, improve regularity assumptions also in the cases of bounded sets and prove general exterior determination results. Counterexamples to uniqueness in the inverse fractional conductivity problem with partial data are constructed in another companion work.

Faking Fake News for Real Fake News Detection Propagandaloaded Training Data Generation ; Despite recent advances in detecting fake news generated by neural models, their results are not readily applicable to effective detection of humanwritten disinformation. What limits the successful transfer between them is the sizable gap between machinegenerated fake news and humanauthored ones, including the notable differences in terms of style and underlying intent. With this in mind, we propose a novel framework for generating training examples that are informed by the known styles and strategies of humanauthored propaganda. Specifically, we perform selfcritical sequence training guided by natural language inference to ensure the validity of the generated articles, while also incorporating propaganda techniques, such as appeal to authority and loaded language. In particular, we create a new training dataset, PropaNews, with 2,256 examples, which we release for future use. Our experimental results show that fake news detectors trained on PropaNews are better at detecting humanwritten disinformation by 3.62 7.69 F1 score on two public datasets.

A review of Generative Adversarial Networks for Electronic Health Records applications, evaluation measures and data sources ; Electronic Health Records EHRs are a valuable asset to facilitate clinical research and point of care applications; however, many challenges such as data privacy concerns impede its optimal utilization. Deep generative models, particularly, Generative Adversarial Networks GANs show great promise in generating synthetic EHR data by learning underlying data distributions while achieving excellent performance and addressing these challenges. This work aims to review the major developments in various applications of GANs for EHRs and provides an overview of the proposed methodologies. For this purpose, we combine perspectives from healthcare applications and machine learning techniques in terms of source datasets and the fidelity and privacy evaluation of the generated synthetic datasets. We also compile a list of the metrics and datasets used by the reviewed works, which can be utilized as benchmarks for future research in the field. We conclude by discussing challenges in GANs for EHRs development and proposing recommended practices. We hope that this work motivates novel research development directions in the intersection of healthcare and machine learning.

The landscape of polymer quantum cosmology ; We show that the quantization ambiguities of loop quantum cosmology, when considered in wider generality, can be used to produce discretionary dynamical behavior. There is an infinite dimensional space of ambiguities which parallels the infinite list of higher curvature corrections in perturbative quantum gravity. There is however an ensemble of qualitative consequences which are generic in the sense that they are independent of these ambiguities. Among these, one has well defined fundamental dynamics across the big bang, and the existence of extra microscopic quantum degrees of freedom that might be relevant in discussions about unitarity in quantum gravity. We show that in addition to the well known bouncing solutions of the effective equations there are other generic type of solutions for sufficiently soft initial conditions in the matter sector tunneling solutions where the scale factor goes through zero and the spacetime orientation is inverted. We also show that generically, a contracting semiclassical universe branches off at the big bang into a quantum superposition of universes with different quantum numbers. Despite their lack of quantitative predictive power these models offer a fertile playground for the discussion of qualitative and conceptual issues in quantum gravity.

MCoNaLa A Benchmark for Code Generation from Multiple Natural Languages ; While there has been a recent burgeoning of applications at the intersection of natural and programming languages, such as code generation and code summarization, these applications are usually Englishcentric. This creates a barrier for program developers who are not proficient in English. To mitigate this gap in technology development across languages, we propose a multilingual dataset, MCoNaLa, to benchmark code generation from natural language commands extending beyond English. Modeled off of the methodology from the English CodeNatural Language Challenge CoNaLa dataset, we annotated a total of 896 NLcode pairs in three languages Spanish, Japanese, and Russian. We present a quantitative evaluation of performance on the MCoNaLa dataset by testing with stateoftheart code generation systems. While the difficulties vary across these three languages, all systems lag significantly behind their English counterparts, revealing the challenges in adapting code generation to new languages.

Explaining Preferencedriven Schedules the EXPRES Framework ; Scheduling is the task of assigning a set of scarce resources distributed over time to a set of agents, who typically have preferences about the assignments they would like to get. Due to the constrained nature of these problems, satisfying all agents' preferences is often infeasible, which might lead to some agents not being happy with the resulting schedule. Providing explanations has been shown to increase satisfaction and trust in solutions produced by AI tools. However, it is particularly challenging to explain solutions that are influenced by and impact on multiple agents. In this paper we introduce the EXPRES framework, which can explain why a given preference was unsatisfied in a given optimal schedule. The EXPRES framework consists of i an explanation generator that, based on a MixedInteger Linear Programming model, finds the best set of reasons that can explain an unsatisfied preference; and ii an explanation parser, which translates the generated explanations into human interpretable ones. Through simulations, we show that the explanation generator can efficiently scale to large instances. Finally, through a set of user studies within J.P. Morgan, we show that employees preferred the explanations generated by EXPRES over humangenerated ones when considering workforce scheduling scenarios.

Towards Device Efficient Conditional Image Generation ; We present a novel algorithm to reduce tensor compute required by a conditional image generation autoencoder without sacrificing quality of photorealistic image generation. Our method is device agnostic, and can optimize an autoencoder for a given CPUonly, GPU compute devices in about normal time it takes to train an autoencoder on a generic workstation. We achieve this via a twostage novel strategy where, first, we condense the channel weights, such that, as few as possible channels are used. Then, we prune the nearly zeroed out weight activations, and finetune the autoencoder. To maintain image quality, finetuning is done via studentteacher training, where we reuse the condensed autoencoder as the teacher. We show performance gains for various conditional image generation tasks segmentation mask to face images, face images to cartoonization, and finally CycleGANbased model over multiple compute devices. We perform various ablation studies to justify the claims and design choices, and achieve realtime versions of various autoencoders on CPUonly devices while maintaining image quality, thus enabling atscale deployment of such autoencoders.

Music Generation Using an LSTM ; Over the past several years, deep learning for sequence modeling has grown in popularity. To achieve this goal, LSTM network structures have proven to be very useful for making predictions for the next output in a series. For instance, a smartphone predicting the next word of a text message could use an LSTM. We sought to demonstrate an approach of music generation using Recurrent Neural Networks RNN. More specifically, a Long ShortTerm Memory LSTM neural network. Generating music is a notoriously complicated task, whether handmade or generated, as there are a myriad of components involved. Taking this into account, we provide a brief synopsis of the intuition, theory, and application of LSTMs in music generation, develop and present the network we found to best achieve this goal, identify and address issues and challenges faced, and include potential future improvements for our network.

Multitasking Framework for Unsupervised Simple Definition Generation ; The definition generation task can help language learners by providing explanations for unfamiliar words. This task has attracted much attention in recent years. We propose a novel task of Simple Definition Generation SDG to help language learners and low literacy readers. A significant challenge of this task is the lack of learner's dictionaries in many languages, and therefore the lack of data for supervised training. We explore this task and propose a multitasking framework SimpDefiner that only requires a standard dictionary with complex definitions and a corpus containing arbitrary simple texts. We disentangle the complexity factors from the text by carefully designing a parameter sharing scheme between two decoders. By jointly training these components, the framework can generate both complex and simple definitions simultaneously. We demonstrate that the framework can generate relevant, simple definitions for the target words through automatic and manual evaluations on English and Chinese datasets. Our method outperforms the baseline model by a 1.77 SARI score on the English dataset, and raises the proportion of the low level HSK level 13 words in Chinese definitions by 3.87.

MakeAScene SceneBased TexttoImage Generation with Human Priors ; Recent texttoimage generation methods provide a simple yet exciting conversion capability between text and image domains. While these methods have incrementally improved the generated image fidelity and text relevancy, several pivotal gaps remain unanswered, limiting applicability and quality. We propose a novel texttoimage method that addresses these gaps by i enabling a simple control mechanism complementary to text in the form of a scene, ii introducing elements that substantially improve the tokenization process by employing domainspecific knowledge over key image regions faces and salient objects, and iii adapting classifierfree guidance for the transformer use case. Our model achieves stateoftheart FID and human evaluation results, unlocking the ability to generate high fidelity images in a resolution of 512x512 pixels, significantly improving visual quality. Through scene controllability, we introduce several new capabilities i Scene editing, ii text editing with anchor scenes, iii overcoming outofdistribution text prompts, and iv story illustration generation, as demonstrated in the story we wrote.

iPLAN Interactive and Procedural Layout Planning ; Layout design is ubiquitous in many applications, e.g. architectureurban planning, etc, which involves a lengthy iterative design process. Recently, deep learning has been leveraged to automatically generate layouts via image generation, showing a huge potential to free designers from laborious routines. While automatic generation can greatly boost productivity, designer input is undoubtedly crucial. An ideal AIaided design tool should automate repetitive routines, and meanwhile accept human guidance and provide smartproactive suggestions. However, the capability of involving humans into the loop has been largely ignored in existing methods which are mostly endtoend approaches. To this end, we propose a new humanintheloop generative model, iPLAN, which is capable of automatically generating layouts, but also interacting with designers throughout the whole procedure, enabling humans and AI to coevolve a sketchy idea gradually into the final design. iPLAN is evaluated on diverse datasets and compared with existing methods. The results show that iPLAN has high fidelity in producing similar layouts to those from human designers, great flexibility in accepting designer inputs and providing design suggestions accordingly, and strong generalizability when facing unseen design tasks and limited training data.

Towards Implicit TextGuided 3D Shape Generation ; In this work, we explore the challenging task of generating 3D shapes from text. Beyond the existing works, we propose a new approach for textguided 3D shape generation, capable of producing highfidelity shapes with colors that match the given text description. This work has several technical contributions. First, we decouple the shape and color predictions for learning features in both texts and shapes, and propose the wordlevel spatial transformer to correlate word features from text with spatial features from shape. Also, we design a cyclic loss to encourage consistency between text and shape, and introduce the shape IMLE to diversify the generated shapes. Further, we extend the framework to enable textguided shape manipulation. Extensive experiments on the largest existing textshape benchmark manifest the superiority of this work. The code and the models are available at httpsgithub.comliuzhengzheTowardsImplicit TextGuidedShapeGeneration.

Text2LIVE TextDriven Layered Image and Video Editing ; We present a method for zeroshot, textdriven appearance manipulation in natural images and videos. Given an input image or video and a target text prompt, our goal is to edit the appearance of existing objects e.g., object's texture or augment the scene with visual effects e.g., smoke, fire in a semantically meaningful manner. We train a generator using an internal dataset of training examples, extracted from a single input image or video and target text prompt, while leveraging an external pretrained CLIP model to establish our losses. Rather than directly generating the edited output, our key idea is to generate an edit layer coloropacity that is composited over the original input. This allows us to constrain the generation process and maintain high fidelity to the original input via novel textdriven losses that are applied directly to the edit layer. Our method neither relies on a pretrained generator nor requires userprovided edit masks. We demonstrate localized, semantic edits on highresolution natural images and videos across a variety of objects and scenes.

Long Video Generation with TimeAgnostic VQGAN and TimeSensitive Transformer ; Videos are created to express emotion, exchange information, and share experiences. Video synthesis has intrigued researchers for a long time. Despite the rapid progress driven by advances in visual synthesis, most existing studies focus on improving the frames' quality and the transitions between them, while little progress has been made in generating longer videos. In this paper, we present a method that builds on 3DVQGAN and transformers to generate videos with thousands of frames. Our evaluation shows that our model trained on 16frame video clips from standard benchmarks such as UCF101, Sky Timelapse, and TaichiHD datasets can generate diverse, coherent, and highquality long videos. We also showcase conditional extensions of our approach for generating meaningful long videos by incorporating temporal information with text and audio. Videos and code can be found at httpssongweige.github.ioprojectstatsindex.html.

Generative Negative Replay for Continual Learning ; Learning continually is a key aspect of intelligence and a necessary ability to solve many reallife problems. One of the most effective strategies to control catastrophic forgetting, the Achilles' heel of continual learning, is storing part of the old data and replaying them interleaved with new experiences also known as the replay approach. Generative replay, which is using generative models to provide replay patterns on demand, is particularly intriguing, however, it was shown to be effective mainly under simplified assumptions, such as simple scenarios and lowdimensional data. In this paper, we show that, while the generated data are usually not able to improve the classification accuracy for the old classes, they can be effective as negative examples or antagonists to better learn the new classes, especially when the learning experiences are small and contain examples of just one or few classes. The proposed approach is validated on complex classincremental and dataincremental continual learning scenarios CORe50 and ImageNet1000 composed of highdimensional data and a large number of training experiences a setup where existing generative replay approaches usually fail.

MultiView Consistent Generative Adversarial Networks for 3Daware Image Synthesis ; 3Daware image synthesis aims to generate images of objects from multiple views by learning a 3D representation. However, one key challenge remains existing approaches lack geometry constraints, hence usually fail to generate multiview consistent images. To address this challenge, we propose MultiView Consistent Generative Adversarial Networks MVCGAN for highquality 3Daware image synthesis with geometry constraints. By leveraging the underlying 3D geometry information of generated images, i.e., depth and camera transformation matrix, we explicitly establish stereo correspondence between views to perform multiview joint optimization. In particular, we enforce the photometric consistency between pairs of views and integrate a stereo mixup mechanism into the training process, encouraging the model to reason about the correct 3D shape. Besides, we design a twostage training strategy with featurelevel multiview joint optimization to improve the image quality. Extensive experiments on three datasets demonstrate that MVCGAN achieves the stateoftheart performance for 3Daware image synthesis.

Generalized Lindblad Master Equation for MeasurementInduced Phase Transition ; The measurementinduced phase transition MIPT occurs when the system is evolving under unitary evolution together with local measurements followed by postselection. We propose a generalized version of the Lindblad master equation as a continuous equation, to describe the dynamics of second R'enyi entropy in the MIPT. This generalized Lindblad equation explicitly takes into account the postselection in the MIPT, which is realized as the EinsteinPodolskyRosen EPR state projection in the equation. Also, this generalized Lindblad equation preserves the Hermitian, unit trace, and positive definiteness of the density matrix. We further use the hardcore BoseHubbard model as a concrete example to numerically confirm that our generalized Lindblad equation is applicable to describing the MIPT.

GANimator Neural Motion Synthesis from a Single Sequence ; We present GANimator, a generative model that learns to synthesize novel motions from a single, short motion sequence. GANimator generates motions that resemble the core elements of the original motion, while simultaneously synthesizing novel and diverse movements. Existing datadriven techniques for motion synthesis require a large motion dataset which contains the desired and specific skeletal structure. By contrast, GANimator only requires training on a single motion sequence, enabling novel motion synthesis for a variety of skeletal structures e.g., bipeds, quadropeds, hexapeds, and more. Our framework contains a series of generative and adversarial neural networks, each responsible for generating motions in a specific frame rate. The framework progressively learns to synthesize motion from random noise, enabling hierarchical control over the generated motion content across varying levels of detail. We show a number of applications, including crowd simulation, keyframe editing, style transfer, and interactive control, which all learn from a single input sequence. Code and data for this paper are at httpspeizhuoli.github.ioganimator.

CORWA A CitationOriented Related Work Annotation Dataset ; Academic research is an exploratory activity to discover new solutions to problems. By this nature, academic research works perform literature reviews to distinguish their novelties from prior work. In natural language processing, this literature review is usually conducted under the Related Work section. The task of related work generation aims to automatically generate the related work section given the rest of the research paper and a list of papers to cite. Prior work on this task has focused on the sentence as the basic unit of generation, neglecting the fact that related work sections consist of variable length text fragments derived from different information sources. As a first step toward a linguisticallymotivated related work generation framework, we present a Citation Oriented Related Work Annotation CORWA dataset that labels different types of citation text fragments from different information sources. We train a strong baseline model that automatically tags the CORWA labels on massive unlabeled related work section texts. We further suggest a novel framework for humanintheloop, iterative, abstractive related work generation.

Investigating Generalization by Controlling Normalized Margin ; Weight norm w and margin gamma participate in learning theory via the normalized margin gammaw. Since standard neural net optimizers do not control normalized margin, it is hard to test whether this quantity causally relates to generalization. This paper designs a series of experimental studies that explicitly control normalized margin and thereby tackle two central questions. First does normalized margin always have a causal effect on generalization The paper finds that no networks can be produced where normalized margin has seemingly no relationship with generalization, counter to the theory of Bartlett et al. 2017. Second does normalized margin ever have a causal effect on generalization The paper finds that yes in a standard training setup, test performance closely tracks normalized margin. The paper suggests a Gaussian process model as a promising explanation for this behavior.

ACHILLES A novel event generator for electron and neutrinonucleus scattering ; We present a novel leptonnucleus event generator ACHILLES, A CHIcagoLand Lepton Event Simulator. The generator factorizes the primary interaction from the propagation of hadrons in the nucleus, which allows for a great deal of modularity, facilitating further improvements and interfaces with existing codes. We validate our generator against high quality electroncarbon scattering data in the quasielastic regime, including the recent CLASe4v reanalysis of existing data. We find good agreement in both inclusive and exclusive distributions. By varying the assumptions on the propagation of knocked out nucleons throughout the nucleus, we estimate a component of theoretical uncertainties. We also propose novel observables that will allow for further testing of leptonnucleus scattering models. ACHILLES is readily extendable to generate neutrinonucleus scattering events.

Talking Face Generation with Multilingual TTS ; In this work, we propose a joint system combining a talking face generation system with a texttospeech system that can generate multilingual talking face videos from only the text input. Our system can synthesize natural multilingual speeches while maintaining the vocal identity of the speaker, as well as lip movements synchronized to the synthesized speech. We demonstrate the generalization capabilities of our system by selecting four languages Korean, English, Japanese, and Chinese each from a different language family. We also compare the outputs of our talking face generation model to outputs of a prior work that claims multilingual support. For our demo, we add a translation API to the preprocessing stage and present it in the form of a neural dubber so that users can utilize the multilingual property of our system more easily.

Highorder harmonic generations in tilted Weyl semimetals ; We investigate highorder harmonic generations HHGs under the comparison of Weyl cones in two types. Due to the hyperboloidal electron pocket structure, strong noncentrosymmetrical generations in high orders are observed around a single typeII Weyl point, especially at frequency zero. Such remarkable DC signal is proved to have attributions from the intraband transition after spectral decomposition. Under weak pulse electric field , the linear optical response of a nontilted Weyl cone is consistent with the Kubo theory. With more numerical simulations, we conclude the nonzero chemical potential can enhance the evenorder generations, from the slightly tilted system to the overtilted systems. In consideration of dynamical symmetries, typeI and II Weyl cones also show different selective responses under the circularly polarized light. Finally, using a more realistic model containing two pairs of Weyl points, we demonstrate the paired Weyl points with opposite chirality could suppress the overall evenorder generations.

Understanding Metrics for Paraphrasing ; Paraphrase generation is a difficult problem. This is not only because of the limitations in text generation capabilities but also due that to the lack of a proper definition of what qualifies as a paraphrase and corresponding metrics to measure how good it is. Metrics for evaluation of paraphrasing quality is an on going research problem. Most of the existing metrics in use having been borrowed from other tasks do not capture the complete essence of a good paraphrase, and often fail at borderlinecases. In this work, we propose a novel metric ROUGEP to measure the quality of paraphrases along the dimensions of adequacy, novelty and fluency. We also provide empirical evidence to show that the current natural language generation metrics are insufficient to measure these desired properties of a good paraphrase. We look at paraphrase model finetuning and generation from the lens of metrics to gain a deeper understanding of what it takes to generate and evaluate a good paraphrase.

The MARTY user interface for the calculation of general Wilson coefficients ; The calculation of oneloop Wilson coefficients for general Beyond the Standard Model BSM scenarios is a technical challenge often addressed by doing long and error prone analytical calculations by hand. Several software programs already provide squared amplitude calculations at the looplevel, but few of them are also able to derive general looplevel Wilson coefficients necessary e.g. for the study of quark decays in flavor physics. MARTY, a computer program that automates treelevel and oneloop perturbative calculations for general BSM scenarios can in particular be used to obtain such Wilson coefficients. We present in details the simple user interface allowing to derive common Wilson coefficients in MARTY, and the most general use case of MARTY to extract the coefficient of any effective operator.

Neutrino Physics in TeV Scale Gravity Theories ; In this paper, the general features of the neutrino sector in TeV scale quantum gravity theories, such as ADD and Many Species Theory, is investigated. This class of theories has an inherent way to generate small neutrino masses. After reviewing this mechanism it is generalized to a realistic threeflavour case. Furthermore, a procedure is presented how to diagonalize a mass matrix which is generated by this class of theories and how one can find the Standard Model flavour eigenstates. The developed general approach is applied to two specific scenarios within ADD and Many Species Theory and possible effects on neutrino oscillations and on unitarity of the lepton mixing matrix are calculated. Finally, a short overview of phenomenology which can be potentially testable by the nowadays neutrino experiments is presented.

Dimension Independent Generalization of DPSGD for Overparameterized Smooth Convex Optimization ; This paper considers the generalization performance of differentially private convex learning. We demonstrate that the convergence analysis of Langevin algorithms can be used to obtain new generalization bounds with differential privacy guarantees for DPSGD. More specifically, by using some recently obtained dimensionindependent convergence results for stochastic Langevin algorithms with convex objective functions, we obtain On14 privacy guarantees for DPSGD with the optimal excess generalization error of tildeOn12 for certain classes of overparameterized smooth convex optimization problems. This improves previous DPSGD results for such problems that contain explicit dimension dependencies, so that the resulting generalization bounds become unsuitable for overparameterized models used in practical applications.

Open ERP System Data For Occupational Fraud Detection ; Recent estimates report that companies lose 5 of their revenue to occupational fraud. Since most mediumsized and large companies employ Enterprise Resource Planning ERP systems to track vast amounts of information regarding their business process, researchers have in the past shown interest in automatically detecting fraud through ERP system data. Current research in this area, however, is hindered by the fact that ERP system data is not publicly available for the development and comparison of fraud detection methods. We therefore endeavour to generate public ERP system data that includes both normal business operation and fraud. We propose a strategy for generating ERP system data through a serious game, model a variety of fraud scenarios in cooperation with auditing experts, and generate data from a simulated maketostock production company with multiple research participants. We aggregate the generated data into ready to used datasets for fraud detection in ERP systems, and supply both the raw and aggregated data to the general public to allow for open development and comparison of fraud detection approaches on ERP system data.

Analysis of function approximation and stability of general DNNs in directed acyclic graphs using unrectifying analysis ; A general lack of understanding pertaining to deep feedforward neural networks DNNs can be attributed partly to a lack of tools with which to analyze the composition of nonlinear functions, and partly to a lack of mathematical models applicable to the diversity of DNN architectures. In this paper, we made a number of basic assumptions pertaining to activation functions, nonlinear transformations, and DNN architectures in order to use the unrectifying method to analyze DNNs via directed acyclic graphs DAGs. DNNs that satisfy these assumptions are referred to as general DNNs. Our construction of an analytic graph was based on an axiomatic method in which DAGs are built from the bottomup through the application of atomic operations to basic elements in accordance with regulatory rules. This approach allows us to derive the properties of general DNNs via mathematical induction. We show that using the proposed approach, some properties hold true for general DNNs can be derived. This analysis advances our understanding of network functions and could promote further theoretical insights if the host of analytical tools for graphs can be leveraged.

Doubly Reparameterized Importance Weighted Structure Learning for Scene Graph Generation ; As a structured prediction task, scene graph generation, given an input image, aims to explicitly model objects and their relationships by constructing a visuallygrounded scene graph. In the current literature, such task is universally solved via a message passing neural network based mean field variational Bayesian methodology. The classical loose evidence lower bound is generally chosen as the variational inference objective, which could induce oversimplified variational approximation and thus underestimate the underlying complex posterior. In this paper, we propose a novel doubly reparameterized importance weighted structure learning method, which employs a tighter importance weighted lower bound as the variational inference objective. It is computed from multiple samples drawn from a reparameterizable GumbelSoftmax sampler and the resulting constrained variational inference task is solved by a generic entropic mirror descent algorithm. The resulting doubly reparameterized gradient estimator reduces the variance of the corresponding derivatives with a beneficial impact on learning. The proposed method achieves the stateoftheart performance on various popular scene graph generation benchmarks.

3DAware Video Generation ; Generative models have emerged as an essential building block for many image synthesis and editing tasks. Recent advances in this field have also enabled highquality 3D or video content to be generated that exhibits either multiview or temporal consistency. With our work, we explore 4D generative adversarial networks GANs that learn unconditional generation of 3Daware videos. By combining neural implicit representations with timeaware discriminator, we develop a GAN framework that synthesizes 3D video supervised only with monocular videos. We show that our method learns a rich embedding of decomposable 3D structures and motions that enables new visual effects of spatiotemporal renderings while producing imagery with quality comparable to that of existing 3D or video GANs.

Personalized Showcases Generating MultiModal Explanations for Recommendations ; Existing explanation models generate only text for recommendations but still struggle to produce diverse contents. In this paper, to further enrich explanations, we propose a new task named personalized showcases, in which we provide both textual and visual information to explain our recommendations. Specifically, we first select a personalized image set that is the most relevant to a user's interest toward a recommended item. Then, natural language explanations are generated accordingly given our selected images. For this new task, we collect a largescale dataset from Google Local i.e.,maps and construct a highquality subset for generating multimodal explanations. We propose a personalized multimodal framework which can generate diverse and visuallyaligned explanations via contrastive learning. Experiments show that our framework benefits from different modalities as inputs, and is able to produce more diverse and expressive explanations compared to previous methods on a variety of evaluation metrics.

PGMG A PharmacophoreGuided Deep Learning Approach for Bioactive Molecular Generation ; The rational design of novel molecules with desired bioactivity is a critical but challenging task in drug discovery, especially when treating a novel target family or understudied targets. Here, we propose PGMG, a pharmacophoreguided deep learning approach for bioactivate molecule generation. Through the guidance of pharmacophore, PGMG provides a flexible strategy to generate bioactive molecules with structural diversity in various scenarios using a trained variational autoencoder. We show that PGMG can generate molecules matching given pharmacophore models while maintaining a high level of validity, uniqueness, and novelty. In the case studies, we demonstrate the application of PGMG to generate bioactive molecules in ligandbased and structurebased drug de novo design, as well as in lead optimization scenarios. Overall, the flexibility and effectiveness of PGMG make it a useful tool for accelerating the drug discovery process.

SceneAware Prompt for Multimodal Dialogue Understanding and Generation ; This paper introduces the schemes of Team LingJing's experiments in NLPCC2022SharedTask4 Multimodal Dialogue Understanding and Generation MDUG. The MDUG task can be divided into two phases multimodal context understanding and response generation. To fully leverage the visual information for both scene understanding and dialogue generation, we propose the sceneaware prompt for the MDUG task. Specifically, we utilize the multitasking strategy for jointly modelling the scene and session multimodal understanding. The visual captions are adopted to aware the scene information, while the fixedtype templated prompt based on the scene and sessionaware labels are used to further improve the dialogue generation performance. Extensive experimental results show that the proposed method has achieved stateoftheart SOTA performance compared with other competitive methods, where we rank the 1st in all three subtasks in this MDUG competition.

Functional Generalized Empirical Likelihood Estimation for Conditional Moment Restrictions ; Important problems in causal inference, economics, and, more generally, robust machine learning can be expressed as conditional moment restrictions, but estimation becomes challenging as it requires solving a continuum of unconditional moment restrictions. Previous works addressed this problem by extending the generalized method of moments GMM to continuum moment restrictions. In contrast, generalized empirical likelihood GEL provides a more general framework and has been shown to enjoy favorable smallsample properties compared to GMMbased estimators. To benefit from recent developments in machine learning, we provide a functional reformulation of GEL in which arbitrary models can be leveraged. Motivated by a dual formulation of the resulting infinite dimensional optimization problem, we devise a practical method and explore its asymptotic properties. Finally, we provide kernel and neural networkbased implementations of the estimator, which achieve stateoftheart empirical performance on two conditional moment restriction problems.

Generalizing to Unseen Domains with Wasserstein Distributional Robustness under Limited Source Knowledge ; Domain generalization aims at learning a universal model that performs well on unseen target domains, incorporating knowledge from multiple source domains. In this research, we consider the scenario where different domain shifts occur among conditional distributions of different classes across domains. When labeled samples in the source domains are limited, existing approaches are not sufficiently robust. To address this problem, we propose a novel domain generalization framework called Wasserstein Distributionally Robust Domain Generalization WDRDG, inspired by the concept of distributionally robust optimization. We encourage robustness over conditional distributions within classspecific Wasserstein uncertainty sets and optimize the worstcase performance of a classifier over these uncertainty sets. We further develop a testtime adaptation module leveraging optimal transport to quantify the relationship between the unseen target domain and source domains to make adaptive inference for target data. Experiments on the Rotated MNIST, PACS and the VLCS datasets demonstrate that our method could effectively balance the robustness and discriminability in challenging generalization scenarios.

FairFuse Interactive Visual Support for Fair Consensus Ranking ; Fair consensus building combines the preferences of multiple rankers into a single consensus ranking, while ensuring any group defined by a protected attribute such as race or gender is not disadvantaged compared to other groups. Manually generating a fair consensus ranking is timeconsuming and impractical even for a fairly small number of candidates. While algorithmic approaches for auditing and generating fair consensus rankings have been developed, these have not been operationalized in interactive systems. To bridge this gap, we introduce FairFuse, a visualization system for generating, analyzing, and auditing fair consensus rankings. We construct a data model which includes base rankings entered by rankers, augmented with measures of group fairness, and algorithms for generating consensus rankings with varying degrees of fairness. We design novel visualizations that encode these measures in a parallelcoordinates style rank visualization, with interactions for generating and exploring fair consensus rankings. We describe use cases in which FairFuse supports a decisionmaker in ranking scenarios in which fairness is important, and discuss emerging challenges for future efforts supporting fairnessoriented rank analysis. Code and demo videos available at httpsosf.iohd639.

Effect of Instance Normalization on FineGrained Control for SketchBased Face Image Generation ; Sketching is an intuitive and effective way for content creation. While significant progress has been made for photorealistic image generation by using generative adversarial networks, it remains challenging to take a finegrained control on synthetic content. The instance normalization layer, which is widely adopted in existing image translation networks, washes away details in the input sketch and leads to loss of precise control on the desired shape of the generated face images. In this paper, we comprehensively investigate the effect of instance normalization on generating photorealistic face images from handdrawn sketches. We first introduce a visualization approach to analyze the feature embedding for sketches with a group of specific changes. Based on the visual analysis, we modify the instance normalization layers in the baseline image translation model. We elaborate a new set of handdrawn sketches with 11 categories of specially designed changes and conduct extensive experimental analysis. The results and user studies demonstrate that our method markedly improve the quality of synthesized images and the conformance with user intention.

Actionconditioned Ondemand Motion Generation ; We propose a novel framework, OnDemand MOtion Generation ODMO, for generating realistic and diverse longterm 3D human motion sequences conditioned only on action types with an additional capability of customization. ODMO shows improvements over SOTA approaches on all traditional motion evaluation metrics when evaluated on three public datasets HumanAct12, UESTC, and MoCap. Furthermore, we provide both qualitative evaluations and quantitative metrics demonstrating several firstknown customization capabilities afforded by our framework, including mode discovery, interpolation, and trajectory customization. These capabilities significantly widen the spectrum of potential applications of such motion generation models. The novel ondemand generative capabilities are enabled by innovations in both the encoder and decoder architectures i Encoder Utilizing contrastive learning in lowdimensional latent space to create a hierarchical embedding of motion sequences, where not only the codes of different action types form different groups, but within an action type, codes of similar inherent patterns motion styles cluster together, making them readily discoverable; ii Decoder Using a hierarchical decoding strategy where the motion trajectory is reconstructed first and then used to reconstruct the whole motion sequence. Such an architecture enables effective trajectory control. Our code is released on the Github page httpsgithub.comroychowdhuryresearchODMO

A Multifaceted Benchmarking of Synthetic Electronic Health Record Generation Models ; Synthetic health data have the potential to mitigate privacy concerns when sharing data to support biomedical research and the development of innovative healthcare applications. Modern approaches for data generation based on machine learning, generative adversarial networks GAN methods in particular, continue to evolve and demonstrate remarkable potential. Yet there is a lack of a systematic assessment framework to benchmark methods as they emerge and determine which methods are most appropriate for which use cases. In this work, we introduce a generalizable benchmarking framework to appraise key characteristics of synthetic health data with respect to utility and privacy metrics. We apply the framework to evaluate synthetic data generation methods for electronic health records EHRs data from two large academic medical centers with respect to several use cases. The results illustrate that there is a utilityprivacy tradeoff for sharing synthetic EHR data. The results further indicate that no method is unequivocally the best on all criteria in each use case, which makes it evident why synthetic data generation methods need to be assessed in context.

SubjectSpecific Lesion Generation and PseudoHealthy Synthesis for Multiple Sclerosis Brain Images ; Understanding the intensity characteristics of brain lesions is key for defining imagebased biomarkers in neurological studies and for predicting disease burden and outcome. In this work, we present a novel foregroundbased generative method for modelling the local lesion characteristics that can both generate synthetic lesions on healthy images and synthesize subjectspecific pseudohealthy images from pathological images. Furthermore, the proposed method can be used as a data augmentation module to generate synthetic images for training brain image segmentation networks. Experiments on multiple sclerosis MS brain images acquired on magnetic resonance imaging MRI demonstrate that the proposed method can generate highly realistic pseudohealthy and pseudopathological brain images. Data augmentation using the synthetic images improves the brain image segmentation performance compared to traditional data augmentation methods as well as a recent lesionaware data augmentation technique, CarveMix. The code will be released at httpsgithub.comdogabasaranlesionsynthesis.

Generalized JackiwTeitelboim gravity in presence of Bloch branelike models ; We investigate generalized JackiwTeitelboim gravity, coupling the dilaton field with two scalar matter fields. We obtain the equations of motion of the fields and investigate the linear perturbation of the solutions in general. We study two specific situations that allow analytic solutions with topological behavior and check how the dilaton field, the warp factor and Ricci scalar behave. In particular, we have shown how the parameters can be used to modify the structure of the solutions. Moreover, the perturbations are in general described by intricate coupled differential equations, but in some specific cases we could construct the corresponding zero modes analytically.

Generating Pixel Art Character Sprites using GANs ; Iterating on creating pixel art character sprite sheets is essential to the game development process. However, it can take a lot of effort until the final versions containing different poses and animation clips are achieved. This paper investigates using conditional generative adversarial networks to aid the designers in creating such sprite sheets. We propose an architecture based on Pix2Pix to generate images of characters facing a target side e.g., right given sprites of them in a source pose e.g., front. Experiments with small pixel art datasets yielded promising results, resulting in models with varying degrees of generalization, sometimes capable of generating images very close to the ground truth. We analyze the results through visual inspection and quantitatively with FID.

PCC Paraphrasing with Bottomk Sampling and Cyclic Learning for Curriculum Data Augmentation ; Curriculum Data Augmentation CDA improves neural models by presenting synthetic data with increasing difficulties from easy to hard. However, traditional CDA simply treats the ratio of word perturbation as the difficulty measure and goes through the curriculums only once. This paper presents textbfPCC textbfParaphrasing with Bottomk Sampling and textbfCyclic Learning for textbfCurriculum Data Augmentation, a novel CDA framework via paraphrasing, which exploits the textual paraphrase similarity as the curriculum difficulty measure. We propose a curriculumaware paraphrase generation module composed of three units a paraphrase candidate generator with bottomk sampling, a filtering mechanism and a difficulty measure. We also propose a cyclic learning strategy that passes through the curriculums multiple times. The bottomk sampling is proposed to generate superhard instances for the later curriculums. Experimental results on fewshot text classification as well as dialogue generation indicate that PCC surpasses competitive baselines. Human evaluation and extensive case studies indicate that bottomk sampling effectively generates superhard instances, and PCC significantly improves the baseline dialogue agent.

Optimal Storage for Solar Energy SelfSufficiency ; We determine the energy storage needed to achieve self sufficiency to a given reliability as a function of excess capacity in a combined solarenergy generation and storage system. Based on 40 years of solarenergy data for the St. Louis region, we formulate a statistical model that we use to generate synthetic insolation data over millions of years. We use these data to monitor the energy depletion in the storage system near the winter solstice. From this information, we develop explicit formulas for the required storage and the nature of costoptimized system configurations as a function of reliability and the excess of generation capacity. Minimizing the cost of the combined generation and storage system gives the optimal mix of these two constituents. For an annual failure rate of less than 3, it is sufficient to have a solar generation capacity that slightly exceeds the daily electrical load at the winter solstice, together with a few days of storage.

INTERACTION A Generative XAI Framework for Natural Language Inference Explanations ; XAI with natural language processing aims to produce humanreadable explanations as evidence for AI decisionmaking, which addresses explainability and transparency. However, from an HCI perspective, the current approaches only focus on delivering a single explanation, which fails to account for the diversity of human thoughts and experiences in language. This paper thus addresses this gap, by proposing a generative XAI framework, INTERACTION explaIn aNd predicT thEn queRy with contextuAl CondiTional varIational autOeNcoder. Our novel framework presents explanation in two steps step one Explanation and Label Prediction; and step two Diverse Evidence Generation. We conduct intensive experiments with the Transformer architecture on a benchmark dataset, eSNLI. Our method achieves competitive or better performance against stateoftheart baseline models on explanation generation up to 4.7 gain in BLEU and prediction up to 4.4 gain in accuracy in step one; it can also generate multiple diverse explanations in step two.

FiniteCliquewidth Sets of Existential Rules Toward a General Criterion for Decidable yet Highly Expressive Querying ; In our pursuit of generic criteria for decidable ontologybased querying, we introduce 'finitecliquewidth sets' FCS of existential rules, a modeltheoretically defined class of rule sets, inspired by the cliquewidth measure from graph theory. By a generic argument, we show that FCS ensures decidability of entailment for a sizable class of queries dubbed 'DaMSOQs' subsuming conjunctive queries CQs. The FCS class properly generalizes the class of finiteexpansion sets FES, and for signatures of arity at most 2, the class of boundedtreewidth sets BTS. For higher arities, BTS is only indirectly subsumed by FCS by means of reification. Despite the generality of FCS, we provide a rule set with decidable CQ entailment by virtue of firstorderrewritability that falls outside FCS, thus demonstrating the incomparability of FCS and the class of finiteunification sets FUS. In spite of this, we show that if we restrict ourselves to singleheaded rule sets over signatures of arity at most 2, then FCS subsumes FUS.

Improved SensorBased Animal Behavior Classification Performance through Conditional Generative Adversarial Network ; Many activity classifications segments data into fixed window size for feature extraction and classification. However, animal behaviors have various durations that do not match the predetermined window size. The dense labeling and dense prediction methods address this limitation by predicting labels for every point. Thus, by tracing the starting and ending points, we could know the time location and duration of all occurring activities. Still, the dense prediction could be noisy with misalignments problems. We modified the UNet and Conditional Generative Adversarial Network cGAN with customized loss functions as a training strategy to reduce fragmentation and other misalignments. In cGAN, the discriminator and generator trained against each other like an adversarial competition. The generator produces dense predictions. The discriminator works as a highlevel consistency check, in our case, pushing the generator to predict activities with reasonable duration. The model trained with cGAN shows better or comparable performance in the cow, pig, and UCI HAPT dataset. The cGANtrained modified UNet improved from 92.17 to 94.66 for the UCI HAPT dataset and from 90.85 to 93.18 for pig data compared to previous dense prediction work.

Lightweight LongRange Generative Adversarial Networks ; In this paper, we introduce novel lightweight generative adversarial networks, which can effectively capture longrange dependencies in the image generation process, and produce highquality results with a much simpler architecture. To achieve this, we first introduce a longrange module, allowing the network to dynamically adjust the number of focused sampling pixels and to also augment sampling locations. Thus, it can break the limitation of the fixed geometric structure of the convolution operator, and capture longrange dependencies in both spatial and channelwise directions. Also, the proposed longrange module can highlight negative relations between pixels, working as a regularization to stabilize training. Furthermore, we propose a new generation strategy through which we introduce metadata into the image generation process to provide basic information about target images, which can stabilize and speed up the training process. Our novel longrange module only introduces few additional parameters and is easily inserted into existing models to capture longrange dependencies. Extensive experiments demonstrate the competitive performance of our method with a lightweight architecture.

GenLoco Generalized Locomotion Controllers for Quadrupedal Robots ; Recent years have seen a surge in commerciallyavailable and affordable quadrupedal robots, with many of these platforms being actively used in research and industry. As the availability of legged robots grows, so does the need for controllers that enable these robots to perform useful skills. However, most learningbased frameworks for controller development focus on training robotspecific controllers, a process that needs to be repeated for every new robot. In this work, we introduce a framework for training generalized locomotion GenLoco controllers for quadrupedal robots. Our framework synthesizes generalpurpose locomotion controllers that can be deployed on a large variety of quadrupedal robots with similar morphologies. We present a simple but effective morphology randomization method that procedurally generates a diverse set of simulated robots for training. We show that by training a controller on this large set of simulated robots, our models acquire more general control strategies that can be directly transferred to novel simulated and realworld robots with diverse morphologies, which were not observed during training.

Domain Adversarial Training on Conditional Variational AutoEncoder for Controllable Music Generation ; The variational autoencoder has become a leading framework for symbolic music generation, and a popular research direction is to study how to effectively control the generation process. A straightforward way is to control a model using different conditions during inference. However, in music practice, conditions are usually sequential rather than simple categorical labels, involving rich information that overlaps with the learned representation. Consequently, the decoder gets confused about whether to listen to the latent representation or the condition, and sometimes just ignores the condition. To solve this problem, we leverage domain adversarial training to disentangle the representation from condition cues for better control. Specifically, we propose a condition corruption objective that uses the representation to denoise a corrupted condition. Minimized by a discriminator and maximized by the VAE encoder, this objective adversarially induces a conditioninvariant representation. In this paper, we focus on the task of melody harmonization to illustrate our idea, while our methodology can be generalized to other controllable generative tasks. Demos and experiments show that our methodology facilitates not only conditioninvariant representation learning but also higherquality controllability compared to baselines.

Oddparity perturbations in the most general scalarvectortensor theory ; In the context of the most general scalarvectortensor theory, we study the stability of static spherically symmetric black holes under linear oddparity perturbations. We calculate the action to second order in the linear perturbations to derive a master equation for these perturbations. For this general class of models, we obtain the conditions of noghost and Laplacian instability. Then, we study in detail the generalized ReggeWheeler potential of particular cases to find their stability conditions.

PINEAPPLE Personifying INanimate Entities by Acquiring Parallel Personification data for Learning Enhanced generation ; A personification is a figure of speech that endows inanimate entities with properties and actions typically seen as requiring animacy. In this paper, we explore the task of personification generation. To this end, we propose PINEAPPLE Personifying INanimate Entities by Acquiring Parallel Personification data for Learning Enhanced generation. We curate a corpus of personifications called PersonifCorp, together with automatically generated depersonified literalizations of these personifications. We demonstrate the usefulness of this parallel corpus by training a seq2seq model to personify a given literal input. Both automatic and human evaluations show that finetuning with PersonifCorp leads to significant gains in personificationrelated qualities such as animacy and interestingness. A detailed qualitative analysis also highlights key strengths and imperfections of PINEAPPLE over baselines, demonstrating a strong ability to generate diverse and creative personifications that enhance the overall appeal of a sentence.

A Benchmark for Understanding and Generating Dialogue between Characters in Stories ; Many classical fairy tales, fiction, and screenplays leverage dialogue to advance story plots and establish characters. We present the first study to explore whether machines can understand and generate dialogue in stories, which requires capturing traits of different characters and the relationships between them. To this end, we propose two new tasks including Masked Dialogue Generation and Dialogue Speaker Recognition, i.e., generating missing dialogue turns and predicting speakers for specified dialogue turns, respectively. We build a new dataset DialStory, which consists of 105k Chinese stories with a large amount of dialogue weaved into the plots to support the evaluation. We show the difficulty of the proposed tasks by testing existing models with automatic and manual evaluation on DialStory. Furthermore, we propose to learn explicit character representations to improve performance on these tasks. Extensive experiments and case studies show that our approach can generate more coherent and informative dialogue, and achieve higher speaker recognition accuracy than strong baselines.

Machine Learning on generalized Complete Intersection CalabiYau Manifolds ; Generalized Complete Intersection CalabiYau Manifold gCICY is a new construction of CalabiYau manifolds established recently. However, the generation of new gCICYs using standard algebraic method is very laborious. Due to this complexity, the number of gCICYs and their classification still remain unknown. In this paper, we try to make some progress in this direction using neural network. The results showed that our trained models can have a high precision on the existing type 1,1 and type 2,1 gCICYs in the literature. Moreover, They can achieve a 97 precision in predicting new gCICY which is generated differently from those used for training and testing. This shows that machine learning could be an effective method to classify and generate new gCICY.