Datasets:

sswt

/

arxiv_sample_50K

like 0

Constructive expansion for quartic vector fields theories. I. Low dimensions ; This paper is the first of a series aiming at proving rigorously the analyticity and the Borel summability of generic quartic bosonic and fermionic vector models generalizing the ON vector model in diverse dimensions. Both nonrelativistic Schrodinger and relativistic KleinGordon and Dirac kinetic terms are considered. The 4tensor defining the interactions is constant but otherwise arbitrary, up to the symmetries imposed by the statistics of the field. In this paper, we focus on models of low dimensions bosons and fermions for d 0, 1, and relativistic bosons for d 2. Moreover, we investigate the large N and massless limits along with quenching for fermions in d 1. These results are established using the loop vertex expansion LVE and have applications in different fields, including data sciences, condensed matter and string field theory. In particular, this establishes the Borel summability of the SYK model both at finite and large N.

Activation of nonlocality in bound entanglement ; We discuss the relation between entanglement and nonlocality in the hidden nonlocality scenario. Hidden nonlocality signifies nonlocality that can be activated by applying local filters to a particular state that admits a local hiddenvariable model in the Bell scenario. We present a fullybiseparable threequbit bound entangled state with a local model for the most general nonsequential measurements. This proves for the first time that bound entangled states can admit a local model for general measurements. We furthermore show that the local model breaks down when suitable local filters are applied. Our results demonstrate the first example of activation of nonlocality in bound entanglement. Hence, we show that genuine hidden nonlocality does not imply entanglement distillability.

Collapsing behavior of Ricciflat Kahler metrics and long time solutions of the KahlerRicci flow ; We prove a uniform diameter bound for long time solutions of the normalized KahlerRicci flow on an ndimensional projective manifold X with semiample canonical bundle under the assumption that the Ricci curvature is uniformly bounded for all time in a fixed domain containing a fibre of X over its canonical model Xcan. This assumption on the Ricci curvature always holds when the Kodaira dimension of X is n, n1 or when the general fibre of X over its canonical model is a complex torus. In particular, the normalized KahlerRicci flow converges in GromovHausdorff topolopy to its canonical model when X has Kodaira dimension 1 with KX being semiample and the general fibre of X over its canonical model being a complex torus. We also prove the GromovHausdorff limit of collapsing Ricciflat Kahler metrics on a holomorphically fibred CalabiYau manifold is unique and is homeomorphic to the metric completion of the corresponding twisted KahlerEinstein metric on the regular part of its base.

Discriminative Online Learning for Fast Video Object Segmentation ; We address the highly challenging problem of video object segmentation. Given only the initial mask, the task is to segment the target in the subsequent frames. In order to effectively handle appearance changes and similar background objects, a robust representation of the target is required. Previous approaches either rely on finetuning a segmentation network on the first frame, or employ generative appearance models. Although partially successful, these methods often suffer from impractically low frame rates or unsatisfactory robustness. We propose a novel approach, based on a dedicated target appearance model that is exclusively learned online to discriminate between the target and background image regions. Importantly, we design a specialized loss and customized optimization techniques to enable highly efficient online training. Our lightweight target model is integrated into a carefully designed segmentation network, trained offline to enhance the predictions generated by the target model. Extensive experiments are performed on three datasets. Our approach achieves an overall score of over 70 on YouTubeVOS, while operating at 25 frames per second.

Examining the Capability of GANs to Replace Real Biomedical Images in Classification Models Training ; In this paper, we explore the possibility of generating artificial biomedical images that can be used as a substitute for real image datasets in applied machine learning tasks. We are focusing on generation of realistic chest Xray images as well as on the lymph node histology images using the two recent GAN architectures including DCGAN and PGGAN. The possibility of the use of artificial images instead of real ones for training machine learning models was examined by benchmark classification tasks being solved using conventional and deep learning methods. In particular, a comparison was made by replacing real images with synthetic ones at the model training stage and comparing the prediction results with the ones obtained while training on the real image data. It was found that the drop of classification accuracy caused by such training data substitution ranged between 2.2 and 3.5 for deep learning models and between 5.5 and 13.25 for conventional methods such as LBP Random Forests.

Inertial particle velocity and distribution in vertical turbulent channel flow a numerical and experimental comparison ; This study is concerned with the statistics of vertical turbulent channel flow laden with inertial particles for two different volume concentrations PhiV 3 times 106 and PhiV 5 times 105 at a Stokes number of St 58.6 based on viscous units. Two independent direct numerical simulation models utilizing the pointparticle approach are compared to recent experimental measurements, where all relevant nondimensional parameters are directly matched. While both numerical models are built on the same general approach, details of the implementations are different, particularly regarding how twoway coupling is represented. At low volume loading, both numerical models are in general agreement with the experimental measurements, with certain exceptions near the walls for the wallnormal particle velocity fluctuations. At high loading, these discrepancies are increased, and it is found that particle clustering is overpredicted in the simulations as compared to the experimental observations. Potential reasons for the discrepancies are discussed. As this study is among the first to perform onetoone comparisons of particleladen flow statistics between numerical models and experiments, it suggests that continued efforts are required to reconcile differences between the observed behavior and numerical predictions.

Listen to the Image ; Visualtoauditory sensory substitution devices can assist the blind in sensing the visual environment by translating the visual information into a sound pattern. To improve the translation quality, the task performances of the blind are usually employed to evaluate different encoding schemes. In contrast to the toilsome humanbased assessment, we argue that machine model can be also developed for evaluation, and more efficient. To this end, we firstly propose two distinct crossmodal perception model w.r.t. the lateblind and congenitallyblind cases, which aim to generate concrete visual contents based on the translated sound. To validate the functionality of proposed models, two novel optimization strategies w.r.t. the primary encoding scheme are presented. Further, we conduct sets of humanbased experiments to evaluate and compare them with the conducted machinebased assessments in the crossmodal generation task. Their highly consistent results w.r.t. different encoding schemes indicate that using machine model to accelerate optimization evaluation and reduce experimental cost is feasible to some extent, which could dramatically promote the upgrading of encoding scheme then help the blind to improve their visual perception ability.

Composite Higgs and Dark Matter Model in SU6SO6 ; We consider a realisation of composite Higgs models in the context of SU6 SO6 symmetry, which features a custodial bitriplet, two Higgs doublets and dark matter candidates. This model can arise from an underlying gaugefermion theory. The general vacuum structure is explored using the top partial compositeness to generate a special vacuum characterised by a single angle aligned with the first Higgs doublet. We present the CP and Dark Matter mathbbZ2 parity in two different pNGB bases and analyse the spectra in the absence of tadpoles and tachyons. For the phenomenology, we discuss the constraints from electroweak precision tests and from a potentially light CPodd singlet other than the Dark Matter in the model.

A General Framework for Edited Video and Raw Video Summarization ; In this paper, we build a general summarization framework for both of edited video and raw video summarization. Overall, our work can be divided into three folds 1 Four models are designed to capture the properties of video summaries, i.e., containing important people and objects importance, representative to the video content representativeness, no similar keyshots diversity and smoothness of the storyline storyness. Specifically, these models are applicable to both edited videos and raw videos. 2 A comprehensive score function is built with the weighted combination of the aforementioned four models. Note that the weights of the four models in the score function, denoted as propertyweight, are learned in a supervised manner. Besides, the propertyweights are learned for edited videos and raw videos, respectively. 3 The training set is constructed with both edited videos and raw videos in order to make up the lack of training data. Particularly, each training video is equipped with a pair of mixingcoefficients which can reduce the structure mess in the training set caused by the rough mixture. We test our framework on three datasets, including edited videos, short raw videos and long raw videos. Experimental results have verified the effectiveness of the proposed framework.

Structural modeling using overlapped group penalties for discovering predictive biomarkers for subgroup analysis ; The identification of predictive biomarkers from a large scale of covariates for subgroup analysis has attracted fundamental attention in medical research. In this article, we propose a generalized penalized regression method with a novel penalty function, for enforcing the hierarchy structure between the prognostic and predictive effects, such that a nonzero predictive effect must induce its ancestor prognostic effects being nonzero in the model. Our method is able to select useful predictive biomarkers by yielding a sparse, interpretable, and predictable model for subgroup analysis, and can deal with different types of response variable such as continuous, categorical, and timetoevent data. We show that our method is asymptotically consistent under some regularized conditions. To minimize the generalized penalized regression model, we propose a novel integrative optimization algorithm by integrating the majorizationminimization and the alternating direction method of multipliers, which is named after textttsmog. The enriched simulation study and real case study demonstrate that our method is very powerful for discovering the true predictive biomarkers and identifying subgroups of patients.

Traveling concentration pulses of bacteria in a generalized KellerSegel model ; We formulate the Smoluchowski equation for a runandtumble particle. It includes the mean tumble rate in a chemical field, for which we derive a Markovian response theory. Using a multipole expansion and a reactiondiffusion equation for the chemoattractant field, we derive a polarization extended model, which also includes the recently discovered angle bias. In the adiabatic limit we recover generalized KellerSegel equations with diffusion and chemotactic coefficients that depend on the microscopic swimming parameters. By requiring the tumble rate to be positive, our model possesses an upper bound of the chemotactic drift velocity, which is no longer singular as in the original KellerSegel equations. Using the KellerSegel model, we present an extensive study of traveling bacterial concentration pulses demonstrating how speed, width, and height of the pulse depend on the microscopic parameters. Most importantly, we discover a maximum number of bacteria that the pulse can sustain the maximum carrying capacity. Finally, we obtain a remarkably good match to experimental results on the traveling bacterial pulse. It does not require a second, signaling chemical field nor a singular chemotactic drift velocity.

Collisional quantum thermometry ; We introduce a general framework for thermometry based on collisional models, where ancillas probe the temperature of the environment through an intermediary system. This allows for the generation of correlated ancillas even if they are initially independent. Using tools from parameter estimation theory, we show through a minimal qubit model that individual ancillas can already outperform the thermal CramerRao bound. In addition, due to the steadystate nature of our model, when measured collectively the ancillas always exhibit superlinear scalings of the Fisher information. This means that even collective measurements on pairs of ancillas will already lead to an advantage. As we find in our qubit model, such a feature may be particularly valuable for weak systemancilla interactions. Our approach sets forth the notion of metrology in a sequential interactions setting, and may inspire further advances in quantum thermometry.

ConstraintAware Neural Networks for Riemann Problems ; Neural networks are increasingly used in complex datadriven simulations as surrogates or for accelerating the computation of classical surrogates. In many applications physical constraints, such as mass or energy conservation, must be satisfied to obtain reliable results. However, standard machine learning algorithms are generally not tailored to respect such constraints. We propose two different strategies to generate constraintaware neural networks. We test their performance in the context of frontcapturing schemes for strongly nonlinear wave motion in compressible fluid flow. Precisely, in this context socalled Riemann problems have to be solved as surrogates. Their solution describes the local dynamics of the captured wave front in numerical simulations. Three model problems are considered a cubic flux model problem, an isothermal twophase flow model, and the Euler equations. We demonstrate that a decrease in the constraint deviation correlates with low discretization errors for all model problems, in addition to the structural advantage of fulfilling the constraint.

A covariant simultaneous action for branes ; A covariant simultaneous action for branes in an arbitrary curved background spacetime is considered. The action depends on a pair of independent field variables, the brane embedding functions, through the canonical momentum of a reparametrization invariant geometric model for the brane, and an auxiliary vector field. The form of the action is analogous to a symplectic potential. Extremization of the simultaneous action produces at once the equations of motion and the Jacobi equations for the brane geometric model, and it also provides a convenient shortcut towards its second variation. In this note, we consider geometric models depending only on the intrinsic geometry of the brane worldvolume, and discuss briefly the generalization to extrinsic geometry dependent models. The approach is illustrated for DiracNambuGoto DNG branes. For a relativistic particle, a simultaneous action was introduced by Bazanski, that served as an inspiration for the present work.

Threedimensional universality class of Ising model with powerlawcorrelated critical disorder ; We use largescale Monte Carlo simulations to test the WeinribHalperin criterion that predicts new universality classes in the presence of sufficiently slowly decaying powerlawcorrelated quenched disorder. While new universality classes are reasonably well established, the predicted exponents are controversial. We propose a method of growing such correlated disorder using the threedimensional Ising model as benchmark systems both for generating disorder and studying the resulting phase transition. Critical equilibrium configurations of a disorderfree system are used to define the twovalue distributed random bonds with a small powerlaw exponent given by the pure Ising exponent. Finitesize scaling analysis shows a new universality class with a single phase transition, but the critical exponents nud1.135, etad0.483 differ significantly from theoretical predictions. We find that depending on details of the disorder generation, disorderaveraged quantities can develop peaks at two temperatures for finite sizes. Finally, a layer model with the two values of bonds spatially separated to halves of the system genuinely has multiple phase transitions and thermodynamic properties can be flexibly tuned by adjusting the model parameters.

WaitFree Universality of Consensus in the Infinite Arrival Model ; In classical asynchronous distributed systems composed of a fixed number n of processes where some proportion may fail by crashing, many objects do not have a waitfree linearizable implementation e.g. stacks, queues, etc.. It has been proved that consensus is universal in such systems, which means that this system augmented with consensus objects allows to implement any object that has a sequential specification. To this end, many universal constructions have been proposed in systems augmented with consensus objects or with different equivalent objects or special hardware instructions compareswap, fetchadd, etc.. In this paper, we consider a more general system model called infinite arrival model where infinitely many processes may arrive and leave or crash during a run. We prove that consensus is still universal in this more general model. For that, we propose a universal construction. As a first step we build a weak log for which we propose two implementations using consensus objects for the first and the compareswap special instruction for the other.

Testing the DriftDiffusion Model ; The drift diffusion model DDM is a model of sequential sampling with diffusion Brownian signals, where the decision maker accumulates evidence until the process hits a stopping boundary, and then stops and chooses the alternative that corresponds to that boundary. This model has been widely used in psychology, neuroeconomics, and neuroscience to explain the observed patterns of choice and response times in a range of binary choice decision problems. This paper provides a statistical test for DDM's with general boundaries. We first prove a characterization theorem we find a condition on choice probabilities that is satisfied if and only if the choice probabilities are generated by some DDM. Moreover, we show that the drift and the boundary are uniquely identified. We then use our condition to nonparametrically estimate the drift and the boundary and construct a test statistic.

Generalized groupbased epidemic model for spreading processes on networks GgroupEM ; We develop a generalized groupbased epidemic model GgroupEM framework for any compartmental epidemic model for example; susceptibleinfectedsusceptible, susceptibleinfectedrecovered, susceptibleexposedinfectedrecovered. Here, a group consists of a collection of individual nodes. This model can be used to understand the important dynamic characteristics of a stochastic epidemic spreading over very large complex networks, being informative about the state of groups. Aggregating nodes by groups, the state space becomes smaller than the individualbased approach at the cost of aggregation error, which is strongly bounded by the isoperimetric inequality. We also develop a meanfield approximation of this framework to further reduce the statespace size. Finally, we extend the GgroupEM to multilayer networks. Since the groupbased framework is computationally less expensive and faster than an individualbased framework, then this framework is useful when the simulation time is important.

DeepSketchHair Deep Sketchbased 3D Hair Modeling ; We present sketchhair, a deep learning based tool for interactive modeling of 3D hair from 2D sketches. Given a 3D bust model as reference, our sketching system takes as input a userdrawn sketch consisting of hair contour and a few strokes indicating the hair growing direction within a hair region, and automatically generates a 3D hair model, which matches the input sketch both globally and locally. The key enablers of our system are two carefully designed neural networks, namely, S2ONet, which converts an input sketch to a dense 2D hair orientation field; and O2VNet, which maps the 2D orientation field to a 3D vector field. Our system also supports hair editing with additional sketches in new views. This is enabled by another deep neural network, V2VNet, which updates the 3D vector field with respect to the new sketches. All the three networks are trained with synthetic data generated from a 3D hairstyle database. We demonstrate the effectiveness and expressiveness of our tool using a variety of hairstyles and also compare our method with prior art.

Text Summarization with Pretrained Encoders ; Bidirectional Encoder Representations from Transformers BERT represents the latest incarnation of pretrained language models which have recently advanced a wide range of natural language processing tasks. In this paper, we showcase how BERT can be usefully applied in text summarization and propose a general framework for both extractive and abstractive models. We introduce a novel documentlevel encoder based on BERT which is able to express the semantics of a document and obtain representations for its sentences. Our extractive model is built on top of this encoder by stacking several intersentence Transformer layers. For abstractive summarization, we propose a new finetuning schedule which adopts different optimizers for the encoder and the decoder as a means of alleviating the mismatch between the two the former is pretrained while the latter is not. We also demonstrate that a twostaged finetuning approach can further boost the quality of the generated summaries. Experiments on three datasets show that our model achieves stateoftheart results across the board in both extractive and abstractive settings. Our code is available at httpsgithub.comnlpyangPreSumm

On the Structural Properties of Social Networks and their Measurementcalibrated Synthetic Counterparts ; Datadriven analysis of large social networks has attracted a great deal of research interest. In this paper, we investigate 120 real social networks and their measurementcalibrated synthetic counterparts generated by four wellknown network models. We investigate the structural properties of the networks revealing the correlation profiles of graph metrics across various social domains friendship networks, communication networks, and collaboration networks. We find that the correlation patterns differ across domains. We identify a nonredundant set of metrics to describe social networks. We study which topological characteristics of real networks the models can or cannot capture. We find that the goodnessoffit of the network models depends on the domains. Furthermore, while 2K and stochastic block models lack the capability of generating graphs with large diameter and high clustering coefficient at the same time, they can still be used to mimic social networks relatively efficiently.

Sequential Latent Spaces for Modeling the Intention During Diverse Image Captioning ; Diverse and accurate visionlanguage modeling is an important goal to retain creative freedom and maintain user engagement. However, adequately capturing the intricacies of diversity in language models is challenging. Recent works commonly resort to latent variable models augmented with more or less supervision from object detectors or partofspeech tags. Common to all those methods is the fact that the latent variable either only initializes the sentence generation process or is identical across the steps of generation. Both methods offer no finegrained control. To address this concern, we propose SeqCVAE which learns a latent space for every word position. We encourage this temporal latent space to capture the 'intention' about how to complete the sentence by mimicking a representation which summarizes the future. We illustrate the efficacy of the proposed approach to anticipate the sentence continuation on the challenging MSCOCO dataset, significantly improving diversity metrics compared to baselines while performing on par w.r.t sentence quality.

Learning Similarity Conditions Without Explicit Supervision ; Many realworld tasks require models to compare images along multiple similarity conditions e.g. similarity in color, category or shape. Existing methods often reason about these complex similarity relationships by learning conditionaware embeddings. While such embeddings aid models in learning different notions of similarity, they also limit their capability to generalize to unseen categories since they require explicit labels at test time. To address this deficiency, we propose an approach that jointly learns representations for the different similarity conditions and their contributions as a latent variable without explicit supervision. Comprehensive experiments across three datasets, PolyvoreOutfits, MarylandPolyvore and UTZappos50k, demonstrate the effectiveness of our approach our model outperforms the stateoftheart methods, even those that are strongly supervised with predefined similarity conditions, on fillintheblank, outfit compatibility prediction and triplet prediction tasks. Finally, we show that our model learns different visuallyrelevant semantic subspaces that allow it to generalize well to unseen categories.

When Low Resource NLP Meets Unsupervised Language Model Metapretraining Then Metalearning for Fewshot Text Classification ; Text classification tends to be difficult when data are deficient or when it is required to adapt to unseen classes. In such challenging scenarios, recent studies have often used metalearning to simulate the fewshot task, thus negating implicit common linguistic features across tasks. This paper addresses such problems using metalearning and unsupervised language models. Our approach is based on the insight that having a good generalization from a few examples relies on both a generic model initialization and an effective strategy for adapting this model to newly arising tasks. We show that our approach is not only simple but also produces a stateoftheart performance on a wellstudied sentiment classification dataset. It can thus be further suggested that pretraining could be a promising solution for fewshot learning of many other NLP tasks. The code and the dataset to replicate the experiments are made available at httpsgithub.comzxlzrFewShotNLP.

Combined Task and Action Learning from Human Demonstrations for Mobile Manipulation Applications ; Learning from demonstrations is a promising paradigm for transferring knowledge to robots. However, learning mobile manipulation tasks directly from a human teacher is a complex problem as it requires learning models of both the overall task goal and of the underlying actions. Additionally, learning from a small number of demonstrations often introduces ambiguity with respect to the intention of the teacher, making it challenging to commit to one model for generalizing the task to new settings. In this paper, we present an approach to learning flexible mobile manipulation action models and task goal representations from teacher demonstrations. Our action models enable the robot to consider different likely outcomes of each action and to generate feasible trajectories for achieving them. Accordingly, we leverage a probabilistic framework based on Monte Carlo tree search to compute sequences of feasible actions imitating the teacher intention in new settings without requiring the teacher to specify an explicit goal state. We demonstrate the effectiveness of our approach in complex tasks carried out in realworld settings.

Nonlinear optical effects in inversionsymmetrybreaking superconductors ; We study nonlinear optical responses in superconducting systems with inversion mathcalI symmetrybreaking order parameters. We first show that any superconducting system with mathcalI and timereversal mathcalT symmetries requires an mathcalIbreaking order parameter to support optical transitions between particlehole pair bands. We then use a 1D toy model of an mathcalIbreaking superconductor to numerically calculate linear and nonlinear conductivities, including shift current and second harmonic generations SHG responses. We find that the magnitude of the signal is significantly larger in shift currentSHG response compare to the linear response due to the matrix element effect. We also present various scaling behaviors of the SHG signal, which may be relevant to the recent experimental observation of SHG in cuprates 1. Finally, we confirm the generality of our observations regarding nonlinear responses of mathcalIbreaking superconductors, by analyzing other models including a 1D threeband model and 2D square lattice model.

Higgs masses and couplings in the general 2HDM with unitarity bounds ; We investigate the general two Higgs doublet model imposing both the unitarity conditions and the boundedfrombelow conditions. Both types of conditions restrict the ranges of the parameters of the scalar potential. We study the model in the Higgs basis, i.e. in the basis for the scalar doublets where only one doublet has vacuum expectation value. We use the experimental bounds on the oblique parameter T, to produce scalar particles with masses and cubic and quartic couplings of the Higgs in agreement with the phenomenology. The numerical calculations show that the cubic coupling may be up to 1.6 times larger than in the Standard Model, but it may also be zero or even negative. The quartic coupling is always positive and may be up to four times larger than in the Standard Model.

FewShot Generalization for SingleImage 3D Reconstruction via Priors ; Recent work on singleview 3D reconstruction shows impressive results, but has been restricted to a few fixed categories where extensive training data is available. The problem of generalizing these models to new classes with limited training data is largely open. To address this problem, we present a new model architecture that reframes singleview 3D reconstruction as learnt, category agnostic refinement of a provided, categoryspecific prior. The provided prior shape for a novel class can be obtained from as few as one 3D shape from this class. Our model can start reconstructing objects from the novel class using this prior without seeing any training image for this class and without any retraining. Our model outperforms categoryagnostic baselines and remains competitive with more sophisticated baselines that finetune on the novel categories. Additionally, our network is capable of improving the reconstruction given multiple views despite not being trained on task of multiview reconstruction.

Theory of Optimal Bayesian Feature Filtering ; Optimal Bayesian feature filtering OBF is a supervised screening method designed for biomarker discovery. In this article, we prove two major theoretical properties of OBF. First, optimal Bayesian feature selection under a general family of Bayesian models reduces to filtering if and only if the underlying Bayesian model assumes all features are mutually independent. Therefore, OBF is optimal if and only if one assumes all features are mutually independent, and OBF is the only filter method that is optimal under at least one model in the general Bayesian framework. Second, OBF under independent Gaussian models is consistent under very mild conditions, including cases where the data is nonGaussian with correlated features. This result provides conditions where OBF is guaranteed to identify the correct feature set given enough data, and it justifies the use of OBF in nondesign settings where its assumptions are invalid.

Adversarial Orthogonal Regression Two nonLinear Regressions for Causal Inference ; We propose two nonlinear regression methods, named Adversarial Orthogonal Regression AdOR for additive noise models and Adversarial Orthogonal Structural Equation Model AdOSE for the general case of structural equation models. Both methods try to make the residual of regression independent from regressors while putting no assumption on noise distribution. In both methods, two adversarial networks are trained simultaneously where a regression network outputs predictions and a loss network that estimates mutual information in AdOR and KLdivergence in AdOSE. These methods can be formulated as a minimax twoplayer game; at equilibrium, AdOR finds a deterministic map between inputs and output and estimates mutual information between residual and inputs, while AdOSE estimates a conditional probability distribution of output given inputs. The proposed methods can be used as subroutines to address several learning problems in causality, such as causal direction determination or more generally, causal structure learning and causal model estimation. Synthetic and realworld experiments demonstrate that the proposed methods have a remarkable performance with respect to previous solutions.

Generalized Gibbs Ensemble and stringcharge relations in nested Bethe Ansatz ; The nonequilibrium steady states of integrable models are believed to be described by the Generalized Gibbs Ensemble GGE, which involves all local and quasilocal conserved charges of the model. In this work we investigate integrable lattice models solvable by the nested Bethe Ansatz, with group symmetry SUN, Nge 3. In these models the Bethe Ansatz involves various types of Bethe rapidities corresponding to the nesting procedure, describing the internal degrees of freedom for the excitations. We show that a complete set of charges for the GGE can be obtained from the known fusion hierarchy of transfer matrices. The resulting charges are quasilocal in a certain regime in rapidity space, and they completely fix the rapidity distributions of each string type from each nesting level.

Goodnessoffit tests on manifolds ; We develop a general theory for the goodnessoffit test to nonlinear models. In particular, we assume that the observations are noisy samples of a submanifold defined by a yaosufficiently smooth nonlinear map. The observation noise is additive Gaussian. Our main result shows that the residual of the model fit, by solving a nonlinear leastsquare problem, follows a possibly noncentral chi2 distribution. The parameters of the chi2 distribution are related to the model order and dimension of the problem. We further present a method to select the model orders sequentially. We demonstrate the broad application of the general theory in machine learning and signal processing, including determining the rank of lowrank possibly complexvalued matrices and tensors from noisy, partial, or indirect observations, determining the number of sources in signal demixing, and potential applications in determining the number of hidden nodes in neural networks.

Efficiency Metrics for DataDriven Models A Text Summarization Case Study ; Using datadriven models for solving text summarization or similar tasks has become very common in the last years. Yet most of the studies report basic accuracy scores only, and nothing is known about the ability of the proposed models to improve when trained on more data. In this paper, we define and propose three data efficiency metrics data score efficiency, data time deficiency and overall data efficiency. We also propose a simple scheme that uses those metrics and apply it for a more comprehensive evaluation of popular methods on text summarization and title generation tasks. For the latter task, we process and release a huge collection of 35 million abstracttitle pairs from scientific articles. Our results reveal that among the tested models, the Transformer is the most efficient on both tasks.

An Investigation Into Ondevice Personalization of Endtoend Automatic Speech Recognition Models ; Speakerindependent speech recognition systems trained with data from many users are generally robust against speaker variability and work well for a large population of speakers. However, these systems do not always generalize well for users with very different speech characteristics. This issue can be addressed by building personalized systems that are designed to work well for each specific user. In this paper, we investigate the idea of securely training personalized endtoend speech recognition models on mobile devices so that user data and models never leave the device and are never stored on a server. We study how the mobile training environment impacts performance by simulating ondevice data consumption. We conduct experiments using data collected from speech impaired users for personalization. Our results show that personalization achieved 63.7 relative word error rate reduction when trained in a server environment and 58.1 in a mobile environment. Moving to ondevice personalization resulted in 18.7 performance degradation, in exchange for improved scalability and data privacy. To train the model on device, we split the gradient computation into two and achieved 45 memory reduction at the expense of 42 increase in training time.

Searching for classical geometries in spin foam amplitudes a numerical method ; We develop a numerical method to investigate the semiclassical limit of spin foam amplitudes with many vertices. We test it using the PonzanoRegge model, a spin foam model for threedimensional euclidean gravity, and a transition amplitude with three vertices. We study the summation over bulk spins, and we identify the stationary phase points that dominate it and that correspond to classical geometries. We complement with the numerical analysis of a four vertex transition amplitude and with a modification of the model that includes local curvature. We discuss the generalization of our results to the fourdimensional EPRL spin foam model, and we provide suggestions for new computations.

Summary Level Training of Sentence Rewriting for Abstractive Summarization ; As an attempt to combine extractive and abstractive summarization, Sentence Rewriting models adopt the strategy of extracting salient sentences from a document first and then paraphrasing the selected ones to generate a summary. However, the existing models in this framework mostly rely on sentencelevel rewards or suboptimal labels, causing a mismatch between a training objective and evaluation metric. In this paper, we present a novel training signal that directly maximizes summarylevel ROUGE scores through reinforcement learning. In addition, we incorporate BERT into our model, making good use of its ability on natural language understanding. In extensive experiments, we show that a combination of our proposed model and training procedure obtains new stateoftheart performance on both CNNDaily Mail and New York Times datasets. We also demonstrate that it generalizes better on DUC2002 test set.

Nonsmooth variational regularization for processing manifoldvalued data ; Many methods for processing scalar and vector valued images, volumes and other data in the context of inverse problems are based on variational formulations. Such formulations require appropriate regularization functionals that model expected properties of the object to reconstruct. Prominent examples of regularization functionals in a vectorspace context are the total variation TV and the MumfordShah functional, as well as higherorder schemes such as total generalized variation models. Driven by applications where the signals or data live in nonlinear manifolds, there has been quite some interest in developing analogous methods for nonlinear, manifoldvalued data recently. In this chapter, we consider various variational regularization methods for manifoldvalued data. In particular, we consider TV minimization as well as higher order models such as total generalized variation TGV. Also, we discuss discrete MumfordShah models and related methods for piecewise constant data. We develop discrete energies for denoising and report on algorithmic approaches to minimize them. Further, we also deal with the extension of such methods to incorporate indirect measurement terms, thus addressing the inverse problem setup. Finally, we discuss wavelet sparse regularization for manifoldvalued data.

Fermion Mass and Mixing in a LowScale Seesaw Model based on the S4 Flavor Symmetry ; We construct a lowscale seesaw model to generate the masses of active neutrinos based on S4 flavor symmetry supplemented by the Z2 times Z3 times Z4 times Z14times U1L group, capable of reproducing the low energy Standard model SM fermion flavor data. The masses of the SM fermions and the fermionic mixings parameters are generated from a FroggattNielsen mechanism after the spontaneous breaking of the S4times Z2 times Z3 times Z4 times Z14times U1L group. The obtained values for the physical observables of the quark and lepton sectors are in good agreement with the most recent experimental data. The leptonic Dirac CP violating phase de CP is predicted to be 259.579circ and the predictions for the absolute neutrino masses in the model can also saturate the recent constraints.

Distortion Estimation Through Explicit Modeling of the Refractive Surface ; Precise calibration is a must for high reliance 3D computer vision algorithms. A challenging case is when the camera is behind a protective glass or transparent object due to refraction, the image is heavily distorted; the pinhole camera model alone can not be used and a distortion correction step is required. By directly modeling the geometry of the refractive media, we build the image generation process by tracing individual light rays from the camera to a target. Comparing the generated images to their distorted observed counterparts, we estimate the geometry parameters of the refractive surface via model inversion by employing an RBF neural network. We present an image collection methodology that produces data suited for finding the distortion parameters and test our algorithm on synthetic and realworld data. We analyze the results of the algorithm.

A Closer Look at Domain Shift for Deep Learning in Histopathology ; Domain shift is a significant problem in histopathology. There can be large differences in data characteristics of wholeslide images between medical centers and scanners, making generalization of deep learning to unseen data difficult. To gain a better understanding of the problem, we present a study on convolutional neural networks trained for tumor classification of HE stained wholeslide images. We analyze how augmentation and normalization strategies affect performance and learned representations, and what features a trained model respond to. Most centrally, we present a novel measure for evaluating the distance between domains in the context of the learned representation of a particular model. This measure can reveal how sensitive a model is to domain variations, and can be used to detect new data that a model will have problems generalizing to. The results show how learning is heavily influenced by the preparation of training data, and that the latent representation used to do classification is sensitive to changes in data distribution, especially when training without augmentation or normalization.

Modelbased Statistical Depth with Applications to Functional Data ; Statistical depth, a commonly used analytic tool in nonparametric statistics, has been extensively studied for multivariate and functional observations over the past few decades. Although various forms of depth were introduced, they are mainly procedurebased whose definitions are independent of the generative model for observations. To address this problem, we introduce a generative modelbased approach to define statistical depth for both multivariate and functional data. The proposed modelbased depth framework permits simple computation via Monte Carlo sampling and improves the depth estimation accuracy. When applied to functional data, the proposed depth can capture important features such as continuity, smoothness, or phase variability, depending on the defining criteria. Specifically, we view functional data as realizations from a secondorder stochastic process, and define their depths through the eigensystem of the covariance operator. These new definitions are given through a proper metric related to the reproducing kernel Hilbert space of the covariance operator. We propose efficient algorithms to compute the proposed depths and establish estimation consistency. Through simulations and real data, we demonstrate that the proposed functional depths reveal important statistical information such as those captured by the median and quantiles, and detect outliers.

PowNet a power systems analysis model for largescale waterenergy nexus studies ; PowNet is a free modelling tool for simulating the Unit Commitment Economic Dispatch of largescale power systems. PowNet is specifically conceived for applications in the waterenergy nexus domain, which investigate the impact of water availability on electricity supply. To this purpose, PowNet is equipped with features that guarantee accuracy, reusability, and computational efficiency over large spatial and temporal domains. Specifically, the model i accounts for the technoeconomic constraints of both generating units and transmission networks, ii can be easily coupled with models that estimate the status of generating units as a function of the climatic conditions, and iii explicitly includes importexport nodes, which are often found in crossborder systems. PowNet is implemented in Python and runs with the help of any standard optimization solver e.g., Gurobi, CPLEX. Its functionality is demonstrated on the Cambodian power system.

RecordFlux Formal Message Specification and Generation of Verifiable Binary Parsers ; Various vulnerabilities have been found in message parsers of protocol implementations in the past. Even highly sensitive software components like TLS libraries are affected regularly. Resulting issues range from denialofservice attacks to the extraction of sensitive information. The complexity of protocols and imprecise specifications in natural language are the core reasons for subtle bugs in implementations, which are hard to find. The lack of precise specifications impedes formal verification. In this paper, we propose a model and a corresponding domainspecific language to formally specify message formats of existing realworld binary protocols. A unique feature of the model is the capability to define invariants, which specify relations and dependencies between message fields. Furthermore, the model allows defining the relation of messages between different protocol layers and thus ensures correct interpretation of payload data. We present a technique to derive verifiable parsers based on the model, generate efficient code for their implementation, and automatically prove the absence of runtime errors. Examples of parser specifications for Ethernet and TLS demonstrate the applicability of our approach.

General relativistic polarized radiative transfer with inverse Compton scatterings ; We present tt radpol a numerical scheme for integrating multifrequency polarized radiative transfer equations along rays propagating in a curved spacetime. The scheme includes radiative processes such as synchrotron emission, absorption, Faraday rotation and conversion, and, for the first time, relativistic Compton scatterings including effects of light polarization. The scheme is fully covariant and is applicable to model radiogammaray emission and its polarization from, e.g., relativistic jets and accretion flows onto black holes and other exotic objects described in alternative metric theories and modeled semianalytically or with timedependent magnetohydrodynamical simulations. We perform a few tests to validate the implemented numerical algorithms that handle light polarization in curved spacetime. We demonstrate application of the scheme to model broadband emission spectra from a relativistically hot, geometrically thick coronallike inflow around a supermassive black hole where the disk model is realized in a general relativistic magnetohydrodynamical simulation.

Hilbert Space Fragmentation and AshkinTeller Criticality in Fluctuation Coupled Ising Models ; We discuss the effects of exponential fragmentation of the Hilbert space on phase transitions in the context of coupled ferromagnetic Ising models in arbitrary dimension with special emphasis on the one dimensional case. We show that the dynamics generated by quantum fluctuations is bounded within spatial partitions of the system and weak mixing of these partitions caused by global transverse fields leads to a zero temperature phase with ordering in the local product of both Ising copies but no long range order in either species. This leads to a natural connection with the AshkinTeller universality class for general lattices. We confirm this for the periodic chain using quantum Monte Carlo simulations. We also point out that our treatment provides an explanation for pseudofirst order behavior seen in the Binder cumulants of the classical frustrated J1J2 Ising model and the q4 Potts model in 2D.

Radiative typeI seesaw neutrino masses ; We discuss a radiative typeI seesaw. In these models, the radiative generation of Dirac neutrino masses allows to explain the smallness of the observed neutrino mass scale for rather light righthanded neutrino masses in a typeI seesaw. We first present the general idea in a model independent way. This allows us to estimate the typical scale of righthanded neutrino mass as a function of the number of loops. We then present two example models, one at oneloop and another one at twoloop, in which we discuss neutrino masses and lepton flavour violating constraints in more detail. For the twoloop example, righthanded neutrino masses must lie below 100 GeV, thus making this class of models testable in heavy neutral lepton searches.

Collaborative Behavior Models for Optimized HumanRobot Teamwork ; Effective humanrobot collaboration requires informed anticipation. The robot must anticipate the human's actions, but also react quickly and intuitively when its predictions are wrong. The robot must plan its actions to account for the human's own plan, with the knowledge that the human's behavior will change based on what the robot actually does. This cyclical game of predicting a human's future actions and generating a corresponding motion plan is extremely difficult to model using standard techniques. In this work, we describe a novel Model Predictive Control MPCbased framework for finding optimal trajectories in a collaborative, multiagent setting, in which we simultaneously plan for the robot while predicting the actions of its external collaborators. We use humanrobot handovers to demonstrate that with a strong model of the collaborator, our framework produces fluid, reactive humanrobot interactions in novel, cluttered environments. Our method efficiently generates coordinated trajectories, and achieves a high success rate in handover, even in the presence of significant sensor noise.

Bayesian generalized linear model for over and under dispersed counts ; Bayesian models that can handle both over and under dispersed counts are rare in the literature, perhaps because full probability distributions for dispersed counts are rather difficult to construct. This note takes a first look at Bayesian ConwayMaxwellPoisson generalized linear models that can handle both over and under dispersion yet retain the parsimony and interpretability of classical count regression models. The focus is on providing an explicit demonstration of Bayesian regression inferences for dispersed counts via a MetropolisHastings algorithm. We illustrate the approach on two data analysis examples and demonstrate some favourable frequentist properties via a simulation study.

Variational Tracking and Prediction with Generative Disentangled StateSpace Models ; We address tracking and prediction of multiple moving objects in visual data streams as inference and sampling in a disentangled latent statespace model. By encoding objects separately and including explicit position information in the latent state space, we perform tracking via amortized variational Bayesian inference of the respective latent positions. Inference is implemented in a modular neural framework tailored towards our disentangled latent space. Generative and inference model are jointly learned from observations only. Comparing to related prior work, we empirically show that our Markovian statespace assumption enables faithful and much improved longterm prediction well beyond the training horizon. Further, our inference model correctly decomposes frames into objects, even in the presence of occlusions. Tracking performance is increased significantly over prior art.

Monte Carlo Basin Bifurcation Analysis ; Many highdimensional complex systems exhibit an enormously complex landscape of possible asymptotic states. Here, we present a numerical approach geared towards analyzing such systems. It is situated between the classical analysis with macroscopic order parameters and a more thorough, detailed bifurcation analysis. With our machine learning method, based on random sampling and clustering methods, we are able to characterize the different asymptotic states or classes thereof and even their basins of attraction. In order to do this, suitable, easy to compute, statistics of trajectories with randomly generated initial conditions and parameters are clustered by an algorithm such as DBSCAN. Due to its modular and flexible nature, our method has a wide range of possible applications. Typical applications are oscillator networks, but it is not limited only to ordinary differential equation systems, every complex system yielding trajectories, such as maps or agentbased models, can be analyzed, as we show by applying it the DoddsWatts model, a generalized SIRSmodel. A second order Kuramoto model and a StuartLandau oscillator network, each exhibiting a complex multistable regime, are shown as well. The method is available to use as a package for the Julia language.

Modeling Sequences with Quantum States A Look Under the Hood ; Classical probability distributions on sets of sequences can be modeled using quantum states. Here, we do so with a quantum state that is pure and entangled. Because it is entangled, the reduced densities that describe subsystems also carry information about the complementary subsystem. This is in contrast to the classical marginal distributions on a subsystem in which information about the complementary system has been integrated out and lost. A training algorithm based on the density matrix renormalization group DMRG procedure uses the extra information contained in the reduced densities and organizes it into a tensor network model. An understanding of the extra information contained in the reduced densities allow us to examine the mechanics of this DMRG algorithm and study the generalization error of the resulting model. As an illustration, we work with the evenparity dataset and produce an estimate for the generalization error as a function of the fraction of the dataset used in training.

Unsupervised Context Rewriting for Open Domain Conversation ; Context modeling has a pivotal role in open domain conversation. Existing works either use heuristic methods or jointly learn context modeling and response generation with an encoderdecoder framework. This paper proposes an explicit context rewriting method, which rewrites the last utterance by considering context history. We leverage pseudoparallel data and elaborate a context rewriting network, which is built upon the CopyNet with the reinforcement learning method. The rewritten utterance is beneficial to candidate retrieval, explainable context modeling, as well as enabling to employ a singleturn framework to the multiturn scenario. The empirical results show that our model outperforms baselines in terms of the rewriting quality, the multiturn response generation, and the endtoend retrievalbased chatbots.

Learning to Make Generalizable and Diverse Predictions for Retrosynthesis ; We propose a new model for making generalizable and diverse retrosynthetic reaction predictions. Given a target compound, the task is to predict the likely chemical reactants to produce the target. This generative task can be framed as a sequencetosequence problem by using the SMILES representations of the molecules. Building on top of the popular Transformer architecture, we propose two novel pretraining methods that construct relevant auxiliary tasks plausible reactions for our problem. Furthermore, we incorporate a discrete latent variable model into the architecture to encourage the model to produce a diverse set of alternative predictions. On the 50k subset of reaction examples from the United States patent literature USPTO50k benchmark dataset, our model greatly improves performance over the baseline, while also generating predictions that are more diverse.

Composite Neural Network Theory and Application to PM2.5 Prediction ; This work investigates the framework and performance issues of the composite neural network, which is composed of a collection of pretrained and noninstantiated neural network models connected as a rooted directed acyclic graph for solving complicated applications. A pretrained neural network model is generally well trained, targeted to approximate a specific function. Despite a general belief that a composite neural network may perform better than a single component, the overall performance characteristics are not clear. In this work, we construct the framework of a composite network, and prove that a composite neural network performs better than any of its pretrained components with a high probability bound. In addition, if an extra pretrained component is added to a composite network, with high probability, the overall performance will not be degraded. In the study, we explore a complicated application PM2.5 prediction to illustrate the correctness of the proposed composite network theory. In the empirical evaluations of PM2.5 prediction, the constructed composite neural network models support the proposed theory and perform better than other machine learning models, demonstrate the advantages of the proposed framework.

A general approach to maximise information density in neutron reflectometry analysis ; Neutron and Xray reflectometry are powerful techniques facilitating the study of the structure of interfacial materials. The analysis of these techniques is illposed in nature requiring the application of a modeldependent methods. This can lead to the over and under analysis of experimental data, when too many or too few parameters are allowed to vary in the model. In this work, we outline a robust and generic framework for the determination of the set of free parameters that is capable of maximising the information density of the model. This framework involves the determination of the Bayesian evidence for each permutation of free parameters; and is applied to a simple phospholipid monolayer. We believe this framework should become an important component in reflectometry data analysis, and hope others more regularly consider the relative evidence for their analytical models.

A Stochastic Automata Network Description for Spatial DNAMethylation Models ; DNA methylation is an important biological mechanism to regulate gene expression and control cell development. Mechanistic modeling has become a popular approach to enhance our understanding of the dynamics of methylation pattern formation in living cells. Recent findings suggest that the methylation state of a cytosine base can be influenced by its DNA neighborhood. Therefore, it is necessary to generalize existing mathematical models that consider only one cytosine and its partner on the opposite DNAstrand CpG, in order to include such neighborhood dependencies. One approach is to describe the system as a stochastic automata network SAN with functional transitions. We show that singleCpG models can successfully be generalized to multiple CpGs using the SAN description and verify the results by comparing them to results from extensive MonteCarlo simulations.

Quantile Regression Modelling via Location and Scale Mixtures of Normal Distributions ; We show that the estimating equations for quantile regression can be solved using a simple EM algorithm in which the Mstep is computed via weighted least squares, with weights computed at the Estep as the expectation of independent generalized inverseGaussian variables. We compute the variancecovariance matrix for the quantile regression coefficients using a kernel density estimator that results in more stable standard errors than those produced by existing software. A natural modification of the EM algorithm that involves fitting a linear mixed model at the Mstep extends the methodology to mixed effects quantile regression models. In this case, the fitting method can be justified as a generalized alternating minimization algorithm. Obtaining quantile regression estimates via the weighted least squares method enables model diagnostic techniques similar to the ones used in the linear regression setting. The computational approach is compared with existing software using simulated data, and the methodology is illustrated with several case studies.

FontGAN A Unified Generative Framework for Chinese Character Stylization and Destylization ; Chinese character synthesis involves two related aspects, i.e., style maintenance and content consistency. Although some methods have achieved remarkable success in synthesizing a character with specified style from standard font, how to map characters to a specified style domain without losing their identifiability remains very challenging. In this paper, we propose a novel model named FontGAN, which integrates the character stylization and destylization into a unified framework. In our model, we decouple character images into style representation and content representation, which facilitates more precise control of these two types of variables, thereby improving the quality of the generated results. We also introduce two modules, namely, font consistency module FCM and content prior module CPM. FCM exploits a category guided KullbackLeibler loss to embedding the style representation into different Gaussian distributions. It constrains the characters of the same font in the training set globally. On the other hand, it enables our model to obtain style variables through sampling in testing phase. CPM provides content prior for the model to guide the content encoding process and alleviates the problem of stroke deficiency during destylization. Extensive experimental results on character stylization and destylization have demonstrated the effectiveness of our method.

Qubits as edge state detectors illustration using the SSH model ; As is well known, qubits are the fundamental building blocks of quantum computers, and more generally, of quantum information. A major challenge in the development of quantum devices arises because the information content in any quantum state is rather fragile, as no system is completely isolated from its environment. Generally, such interactions degrade the quantum state, resulting in a loss of information. Topological edge states are promising in this regard because they are in ways more robust against noise and decoherence. But creating and detecting edge states can be challenging. We describe a composite system consisting of a twolevel system the qubit interacting with a finite SuSchriefferHeeger chain a hopping model with alternating hopping parameters attached to an infinite chain. In this model, the dynamics of the qubit changes dramatically depending on whether or not an edge state exists. Thus, the qubit can be used to determine whether or not an edge state exists in this model.

The conditional censored graphical lasso estimator ; In many applied fields, such as genomics, different types of data are collected on the same system, and it is not uncommon that some of these datasets are subject to censoring as a result of the measurement technologies used, such as data generated by polymerase chain reactions and flow cytometer. When the overall objective is that of network inference, at possibly different levels of a system, information coming from different sources andor different steps of the analysis can be integrated into one model with the use of conditional graphical models. In this paper, we develop a doubly penalized inferential procedure for a conditional Gaussian graphical model when data can be subject to censoring. The computational challenges of handling censored data in high dimensionality are met with the development of an efficient ExpectationMaximization algorithm, based on approximate calculations of the moments of truncated Gaussian distributions and on a suitably derived twostep procedure alternating graphical lasso with a novel blockcoordinate multivariate lasso approach. We evaluate the performance of this approach on an extensive simulation study and on gene expression data generated by RTqPCR technologies, where we are able to integrate network inference, differential expression detection and data normalization into one model.

Balancing Multilevel Interactions for Sessionbased Recommendation ; Predicting user actions based on anonymous sessions is a challenge to general recommendation systems because the lack of user profiles heavily limits datadriven models. Recently, sessionbased recommendation methods have achieved remarkable results in dealing with this task. However, the upper bound of performance can still be boosted through the innovative exploration of limited data. In this paper, we propose a novel method, namely Intraand Intersession Interactionaware Graphenhanced Network, to take intersession itemlevel interactions into account. Different from existing intrasession itemlevel interactions and sessionlevel collaborative information, our introduced data represents complex itemlevel interactions between different sessions. For mining the new data without breaking the equilibrium of the model between different interactions, we construct an intrasession graph and an intersession graph for the current session. The former focuses on itemlevel interactions within a single session and the latter models those between items among neighborhood sessions. Then different approaches are employed to encode the information of two graphs according to different structures, and the generated latent vectors are combined to balance the model across different scopes. Experiments on realworld datasets verify that our method outperforms other stateoftheart methods.

Generalized Method of Moments Estimation for Stochastic Models of DNA Methylation Patterns ; With recent advances in sequencing technologies, large amounts of epigenomic data have become available and computational methods are contributing significantly to the progress of epigenetic research. As an orthogonal approach to methods based on machine learning, mechanistic modeling aims at a description of the mechanisms underlying epigenetic changes. Here, we propose an efficient method for parameter estimation for stochastic models that describe the dynamics of DNA methylation patterns over time. Our method is based on the Generalized Method of Moments GMM and gives results with an accuracy similar to that of maximum likelihoodbased estimation approaches. However, in contrast to the latter, the GMM still allows an efficient and accurate calibration of parameters even if the complexity of the model is increased by considering longer methylation patterns. We show the usefulness of our method by applying it to hairpin bisulfite sequencing data from mouse ESCs for varying pattern lengths.

MemoryAugmented Recurrent Neural Networks Can Learn Generalized Dyck Languages ; We introduce three memoryaugmented Recurrent Neural Networks MARNNs and explore their capabilities on a series of simple language modeling tasks whose solutions require stackbased mechanisms. We provide the first demonstration of neural networks recognizing the generalized Dyck languages, which express the core of what it means to be a language with hierarchical structure. Our memoryaugmented architectures are easy to train in an endtoend fashion and can learn the Dyck languages over as many as six parenthesispairs, in addition to two deterministic palindrome languages and the stringreversal transduction task, by emulating pushdown automata. Our experiments highlight the increased modeling capacity of memoryaugmented models over simple RNNs, while inflecting our understanding of the limitations of these models.

Bootstrapping Disjoint Datasets for Multilingual Multimodal Representation Learning ; Recent work has highlighted the advantage of jointly learning grounded sentence representations from multiple languages. However, the data used in these studies has been limited to an aligned scenario the same images annotated with sentences in multiple languages. We focus on the more realistic disjoint scenario in which there is no overlap between the images in multilingual imagecaption datasets. We confirm that training with aligned data results in better grounded sentence representations than training with disjoint data, as measured by imagesentence retrieval performance. In order to close this gap in performance, we propose a pseudopairing method to generate synthetically aligned EnglishGermanimage triplets from the disjoint sets. The method works by first training a model on the disjoint data, and then creating new triples across datasets using sentence similarity under the learned model. Experiments show that pseudopairs improve imagesentence retrieval performance compared to disjoint training, despite requiring no external data or models. However, we do find that using an external machine translation model to generate the synthetic data sets results in better performance.

Increasing Robustness to Spurious Correlations using Forgettable Examples ; Neural NLP models tend to rely on spurious correlations between labels and input features to perform their tasks. Minority examples, i.e., examples that contradict the spurious correlations present in the majority of data points, have been shown to increase the outofdistribution generalization of pretrained language models. In this paper, we first propose using example forgetting to find minority examples without prior knowledge of the spurious correlations present in the dataset. Forgettable examples are instances either learned and then forgotten during training or never learned. We empirically show how these examples are related to minorities in our training sets. Then, we introduce a new approach to robustify models by finetuning our models twice, first on the full training data and second on the minorities only. We obtain substantial improvements in outofdistribution generalization when applying our approach to the MNLI, QQP, and FEVER datasets.

Effect of Vacancy Creation and Annihilation on Grain Boundary Motion ; Interaction of vacancies with grain boundaries GBs is involved in many processes occurring in materials, including radiation damage healing, diffusional creep, and solidstate sintering. We analyze a model describing a set of processes occurring at a GB in the presence of a nonequilibrium, nonhomogeneous vacancy concentration. Such processes include vacancy diffusion toward, away from, and across the GB, vacancy generation and absorption at the GB, and GB migration. Numerical calculations within this model reveal that the coupling among the different processes gives rise to interesting phenomena, such as vacancydriven GB motion and accelerated vacancy generationabsorption due to GB motion. The key combinations of the model parameters that control the kinetic regimes of the vacancyGB interactions are identified via a linear stability analysis. Possible applications and extensions of the model are discussed.e's comments

Conditionally Learn to Pay Attention for Sequential Visual Task ; Sequential visual task usually requires to pay attention to its current interested object conditional on its previous observations. Different from popular soft attention mechanism, we propose a new attention framework by introducing a novel conditional global feature which represents the weak feature descriptor of the current focused object. Specifically, for a standard CNN Convolutional Neural Network pipeline, the convolutional layers with different receptive fields are used to produce the attention maps by measuring how the convolutional features align to the conditional global feature. The conditional global feature can be generated by different recurrent structure according to different visual tasks, such as a simple recurrent neural network for multiple objects recognition, or a moderate complex language model for image caption. Experiments show that our proposed conditional attention model achieves the best performance on the SVHN Street View House Numbers dataset with without extra bounding box; and for image caption, our attention model generates better scores than the popular soft attention model.

WhiteBox Target Attack for EEGBased BCI Regression Problems ; Machine learning has achieved great success in many applications, including electroencephalogram EEG based braincomputer interfaces BCIs. Unfortunately, many machine learning models are vulnerable to adversarial examples, which are crafted by adding deliberately designed perturbations to the original inputs. Many adversarial attack approaches for classification problems have been proposed, but few have considered target adversarial attacks for regression problems. This paper proposes two such approaches. More specifically, we consider whitebox target attacks for regression problems, where we know all information about the regression model to be attacked, and want to design small perturbations to change the regression output by a predetermined amount. Experiments on two BCI regression problems verified that both approaches are effective. Moreover, adversarial examples generated from both approaches are also transferable, which means that we can use adversarial examples generated from one known regression model to attack an unknown regression model, i.e., to perform blackbox attacks. To our knowledge, this is the first study on adversarial attacks for EEGbased BCI regression problems, which calls for more attention on the security of BCI systems.

Unsupervised Medical Image Segmentation with Adversarial Networks From Edge Diagrams to Segmentation Maps ; We develop and approach to unsupervised semantic medical image segmentation that extends previous work with generative adversarial networks. We use existing edge detection methods to construct simple edge diagrams, train a generative model to convert them into synthetic medical images, and construct a dataset of synthetic images with known segmentations using variations on extracted edge diagrams. This synthetic dataset is then used to train a supervised image segmentation model. We test our approach on a clinical dataset of kidney ultrasound images and the benchmark ISIC 2018 skin lesion dataset. We show that our unsupervised approach is more accurate than previous unsupervised methods, and performs reasonably compared to supervised image segmentation models. All code and trained models are available at httpsgithub.comkiretdUnsupervisedMIseg.

Adaptive Basis Construction and Improved Error Estimation for Parametric Nonlinear Dynamical Systems ; An adaptive scheme to generate reducedorder models for parametric nonlinear dynamical systems is proposed. It aims to automatize the PODGreedy algorithm combined with empirical interpolation. At each iteration, it is able to adaptively determine the number of the reduced basis vectors and the number of the interpolation basis vectors for basis construction. The proposed technique is able to derive a suitable match between the reduced basis and the interpolation basis vectors, making the generation of a stable, compact and reliable reducedorder model possible. This is achieved by adaptively adding new basis vectors or removing unnecessary ones, at each iteration of the greedy algorithm. An efficient output error indicator plays a key role in the adaptive scheme. We also propose an improved output error indicator based on previous work. Upon convergence of the PODGreedy algorithm, the new error indicator is shown to be sharper than the existing ones, implicating that a more reliable reducedorder model can be constructed. The proposed method is tested on several nonlinear dynamical systems, namely, the viscous Burgers' equation and two other models from chemical engineering.

A Stable Variational Autoencoder for Text Modelling ; Variational Autoencoder VAE is a powerful method for learning representations of highdimensional data. However, VAEs can suffer from an issue known as latent variable collapse or KL loss vanishing, where the posterior collapses to the prior and the model will ignore the latent codes in generative tasks. Such an issue is particularly prevalent when employing VAERNN architectures for text modelling Bowman et al., 2016. In this paper, we present a simple architecture called holistic regularisation VAE HRVAE, which can effectively avoid latent variable collapse. Compared to existing VAERNN architectures, we show that our model can achieve much more stable training process and can generate text with significantly better quality.

Mark my Word A SequencetoSequence Approach to Definition Modeling ; Defining words in a textual context is a useful task both for practical purposes and for gaining insight into distributed word representations. Building on the distributional hypothesis, we argue here that the most natural formalization of definition modeling is to treat it as a sequencetosequence task, rather than a wordtosequence task given an input sequence with a highlighted word, generate a contextually appropriate definition for it. We implement this approach in a Transformerbased sequencetosequence model. Our proposal allows to train contextualization and definition generation in an endtoend fashion, which is a conceptual improvement over earlier works. We achieve stateoftheart results both in contextual and noncontextual definition modeling.

Mining News Events from Comparable News Corpora A MultiAttribute Proximity Network Modeling Approach ; We present ProxiModel, a novel event mining framework for extracting highquality structured event knowledge from large, redundant, and noisy news data sources. The proposed model differentiates itself from other approaches by modeling both the event correlation within each individual document as well as across the corpus. To facilitate this, we introduce the concept of a proximitynetwork, a novel spaceefficient data structure to facilitate scalable event mining. This proximity network captures the corpuslevel cooccurence statistics for candidate event descriptors, event attributes, as well as their connections. We probabilistically model the proximity network as a generative process with sparsityinducing regularization. This allows us to efficiently and effectively extract highquality and interpretable news events. Experiments on three different news corpora demonstrate that the proposed method is effective and robust at generating highquality event descriptors and attributes. We briefly detail many interesting applications from our proposed framework such as news summarization, event tracking and multidimensional analysis on news. Finally, we explore a case study on visualizing the events for a Japan Tsunami news corpus and demonstrate ProxiModel's ability to automatically summarize emerging news events.

Integrating Threat Modeling and Automated Test Case Generation into Industrialized Software Security Testing ; Industrial Internet of Things IIoT application provide a whole new set of possibilities to drive efficiency of industrial production forward. However, with the higher degree of integration among systems, comes a plethora of newthreats to the latter, as they are not yet designed to be broadly reachable and interoperable. To mitigate these vast amount of new threats, systematic and automated test methods are necessary. This comprehensiveness can be achieved by thorough threat modeling. In order to automate security test, we present an approach to automate the testing process from threat modeling onward, closing the gap between threat modeling and automated test case generation.

Improved Document Modelling with a Neural Discourse Parser ; Despite the success of attentionbased neural models for natural language generation and classification tasks, they are unable to capture the discourse structure of larger documents. We hypothesize that explicit discourse representations have utility for NLP tasks over longer documents or document sequences, which sequencetosequence models are unable to capture. For abstractive summarization, for instance, conventional neural models simply match source documents and the summary in a latent space without explicit representation of text structure or relations. In this paper, we propose to use neural discourse representations obtained from a rhetorical structure theory RST parser to enhance document representations. Specifically, document representations are generated for discourse spans, known as the elementary discourse units EDUs. We empirically investigate the benefit of the proposed approach on two different tasks abstractive summarization and popularity prediction of online petitions. We find that the proposed approach leads to improvements in all cases.

Localizing Occluders with Compositional Convolutional Networks ; Compositional convolutional networks are generative compositional models of neural network features, that achieve state of the art results when classifying partially occluded objects, even when they have not been exposed to occluded objects during training. In this work, we study the performance of CompositionalNets at localizing occluders in images. We show that the original model is not able to localize occluders well. We propose to overcome this limitation by modeling the feature activations as a mixture of vonMisesFisher distributions, which also allows for an endtoend training of CompositionalNets. Our experimental results demonstrate that the proposed extensions increase the model's performance at localizing occluders as well as at classifying partially occluded objects.

Mixtures of multivariate generalized linear models with overlapping clusters ; With the advent of ubiquitous monitoring and measurement protocols, studies have started to focus more and more on complex, multivariate and heterogeneous datasets. In such studies, multivariate response variables are drawn from a heterogeneous population often in the presence of additional covariate information. In order to deal with this intrinsic heterogeneity, regression analyses have to be clustered for different groups of units. Up until now, mixture model approaches assigned units to distinct and nonoverlapping groups. However, not rarely these units exhibit more complex organization and clustering. It is our aim to define a mixture of generalized linear models with overlapping clusters of units. This involves crucially an overlap function, that maps the coefficients of the parent clusters into the the coefficient of the multiple allocation units. We present a computationally efficient MCMC scheme that samples the posterior distribution of the parameters in the model. An example on a twomode network study shows details of the implementation in the case of a multivariate probit regression setting. A simulation study shows the overall performance of the method, whereas an illustration of the voting behaviour on the US supreme court shows how the 9 justices split in two overlapping sets of justices.

Exponential Family Graph Embeddings ; Representing networks in a low dimensional latent space is a crucial task with many interesting applications in graph learning problems, such as link prediction and node classification. A widely applied network representation learning paradigm is based on the combination of random walks for sampling context nodes and the traditional textitSkipGram model to capture centercontext node relationships. In this paper, we emphasize on exponential family distributions to capture rich interaction patterns between nodes in random walk sequences. We introduce the generic textitexponential family graph embedding model, that generalizes random walkbased network representation learning techniques to exponential family conditional distributions. We study three particular instances of this model, analyzing their properties and showing their relationship to existing unsupervised learning models. Our experimental evaluation on realworld datasets demonstrates that the proposed techniques outperform wellknown baseline methods in two downstream machine learning tasks.

Procrustes registration of twodimensional statistical shape models without correspondences ; Statistical shape models are a useful tool in image processing and computer vision. A Procrustres registration of the contours of the same shape is typically perform to align the training samples to learn the statistical shape model. A Procrustes registration between two contours with known correspondences is straightforward. However, these correspondences are not generally available. Manually placed landmarks are often used for correspondence in the design of statistical shape models. However, determining manual landmarks on the contours is timeconsuming and often errorprone. One solution to simultaneously find correspondence and registration is the Iterative Closest Point ICP algorithm. However, ICP requires an initial position of the contours that is close to registration, and it is not robust against outliers. We propose a new strategy, based on Dynamic Time Warping, that efficiently solves the Procrustes registration problem without correspondences. We study the registration performance in a collection of different shape data sets and show that our technique outperforms competing techniques based on the ICP approach. Our strategy is applied to an ensemble of contours of the same shape as an extension of the generalized Procrustes analysis accounting for a lack of correspondence.

SAMSum Corpus A Humanannotated Dialogue Dataset for Abstractive Summarization ; This paper introduces the SAMSum Corpus, a new dataset with abstractive dialogue summaries. We investigate the challenges it poses for automated summarization by testing several models and comparing their results with those obtained on a corpus of news articles. We show that modelgenerated summaries of dialogues achieve higher ROUGE scores than the modelgenerated summaries of news in contrast with human evaluators' judgement. This suggests that a challenging task of abstractive dialogue summarization requires dedicated models and nonstandard quality measures. To our knowledge, our study is the first attempt to introduce a highquality chatdialogues corpus, manually annotated with abstractive summarizations, which can be used by the research community for further studies.

UCNNpred A Universal CNNbased Predictor for Stock Markets ; The performance of financial market prediction systems depends heavily on the quality of features it is using. While researchers have used various techniques for enhancing the stock specific features, less attention has been paid to extracting features that represent general mechanism of financial markets. In this paper, we investigate the importance of extracting such general features in stock market prediction domain and show how it can improve the performance of financial market prediction. We present a framework called UCNNpred, that uses a CNNbased structure. A base model is trained in a specially designed layerwise training procedure over a pool of historical data from many financial markets, in order to extract the common patterns from different markets. Our experiments, in which we have used hundreds of stocks in SP 500 as well as 14 famous indices around the world, show that this model can outperform baseline algorithms when predicting the directional movement of the markets for which it has been trained for. We also show that the base model can be finetuned for predicting new markets and achieve a better performance compared to the state of the art baseline algorithms that focus on constructing marketspecific models from scratch.

Image segmentation of liver stage malaria infection with spatial uncertainty sampling ; Global eradication of malaria depends on the development of drugs effective against the silent, yet obligate liver stage of the disease. The gold standard in drug development remains microscopic imaging of liver stage parasites in in vitro cell culture models. Image analysis presents a major bottleneck in this pipeline since the parasite has significant variability in size, shape, and density in these models. As with other highly variable datasets, traditional segmentation models have poor generalizability as they rely on handcrafted features; thus, manual annotation of liver stage malaria images remains standard. To address this need, we develop a convolutional neural network architecture that utilizes spatial dropout sampling for parasite segmentation and epistemic uncertainty estimation in images of liver stage malaria. Our pipeline produces highprecision segmentations nearly identical to expert annotations, generalizes well on a diverse dataset of liver stage malaria parasites, and promotes independence between learned feature maps to model the uncertainty of generated predictions.

EasytoHard Leveraging Simple Questions for Complex Question Generation ; This paper makes one of the first efforts toward automatically generating complex questions from knowledge graphs. Particularly, we study how to leverage existing simple question datasets for this task, under two separate scenarios using either subquestions of the target complex questions, or distantly related pseudo subquestions when the former are unavailable. First, a competitive base model named CoG2Q is designed to map complex query qraphs to natural language questions. Afterwards, we propose two extension models, namely CoGSub2Q and CoGSubm2Q, respectively for the above two scenarios. The former encodes and copies from a subquestion, while the latter further scores and aggregates multiple pseudo subquestions. Experiment results show that the extension models significantly outperform not only base CoG2Q, but also its augmented variant using simple questions as additional training examples. This demonstrates the importance of instancelevel connections between simple and corresponding complex questions, which may be underexploited by straightforward data augmentation of CoG2Q that builds modellevel connections through learned parameters.

Supersymmetric Hyperbolic models and Decay of Correlations in Two Dimensions ; In this paper we study a family of nonlinear sigmamodels in which the target space is the super manifold H22N. These models generalize Zirnbauer's H22 nonlinear sigmamodel which has a number of special features for which we find analogs in the general case. For example, by supersymmetric localization, the partition function of the H22 model is a constants independent of the coupling constants. Here we show that for the H22N model, the partition function is a multivariate polynomial of degree N1, increasing in each variable. We use these facts to provide estimates on the Fourier and Laplace transforms of the 'tfield' when these models are specialized to mathbbZ2. From the bounds, we conclude the tfield exhibits polynomial decay of correlations and has fluctuations which are at least those of a massless free field.

Attending Form and Context to Generate Specialized OutofVocabularyWords Representations ; We propose a new contextualcompositional neural network layer that handles outofvocabulary OOV words in natural language processing NLP tagging tasks. This layer consists of a model that attends to both the character sequence and the context in which the OOV words appear. We show that our model learns to generate taskspecific textitand sentencedependent OOV word representations without the need for pretraining on an embedding table, unlike previous attempts. We insert our layer in the stateoftheart tagging model of citetplank2016multilingual and thoroughly evaluate its contribution on 23 different languages on the task of jointly tagging partofspeech and morphosyntactic attributes. Our OOV handling method successfully improves performances of this model on every language but one to achieve a new stateoftheart on the Universal Dependencies Dataset 1.4.

Magnetically charged black holes from nonlinear electrodynamics and the Event Horizon Telescope ; Nonlinear electrodynamics NLED theories are wellmotivated extensions of QED in the strong field regime, and have long been studied in the search for regular black hole BH solutions. We consider two wellstudied and wellmotivated NLED models coupled to General Relativity the EulerHeisenberg model and the Bronnikov model. After carefully accounting for the effective geometry induced by the NLED corrections, we determine the shadows of BHs within these two models. We then compare these to the shadow of the supermassive BH M87 recently imaged by the Event Horizon Telescope collaboration. In doing so, we are able to extract upper limits on the black hole magnetic charge, thus providing novel constraints on fundamental physics from this new extraordinary probe.

CPGAN FullSpectrum ContentParsing Generative Adversarial Networks for TexttoImage Synthesis ; Typical methods for texttoimage synthesis seek to design effective generative architecture to model the texttoimage mapping directly. It is fairly arduous due to the crossmodality translation. In this paper we circumvent this problem by focusing on parsing the content of both the input text and the synthesized image thoroughly to model the texttoimage consistency in the semantic level. Particularly, we design a memory structure to parse the textual content by exploring semantic correspondence between each word in the vocabulary to its various visual contexts across relevant images during text encoding. Meanwhile, the synthesized image is parsed to learn its semantics in an objectaware manner. Moreover, we customize a conditional discriminator to model the finegrained correlations between words and image subregions to push for the textimage semantic alignment. Extensive experiments on COCO dataset manifest that our model advances the stateoftheart performance significantly from 35.69 to 52.73 in Inception Score.

Dense networks with scalefree feature ; While previous works have shown that an overwhelming number of scalefree networks are sparse, there still exist some realworld networks including social networks, urban networks, information networks, which are by observation dense. In this paper, we propose a novel framework for generating scalefree graphs with dense feature using two simple yet helpful operations, firstorder subdivision and Lineoperation. From the theoretical point of view, our method can be used not only to produce desired scalefree graphs with density feature, i.e. powerlaw exponent gamma falling into the interval 1gammaleq2, but also to establish many other unexpected networked models, for instance, powerlaw models having large diameter. In addition, the networked models generated upon our framework show especially assortative structure. That is, their own Pearson correlation coefficients are able to achieve the theoretical upper bound. Last but not the least, we find the sizes of community in the proposed models to follow powerlaw in form with respect to modularity maximization.

Laurent polynomial LandauGinzburg models for cominuscule homogeneous spaces ; In this article we construct Laurent polynomial LandauGinzburg models for cominuscule homogeneous spaces. These Laurent polynomial potentials are defined on a particular algebraic torus inside the Lietheoretic mirror model constructed for arbitrary homogeneous spaces in arXivmath0511124. The Laurent polynomial takes a similar shape to the one given in arXivalggeom9603021 for projective complete intersections, i.e. it is the sum of the toric coordinates plus a quantum term. We also give a general enumeration method for the summands in the quantum term of the potential in terms of the quiver introduced in arXivmath0607492, associated to the Langlands dual homogeneous space. This enumeration method generalizes the use of Young diagrams for Grassmannians and Lagrangian Grassmannians and can be defined typeindependently. The obtained Laurent polynomials coincide with the results obtained so far in arXiv1404.4844 and arXiv1304.4958 for quadrics and Lagrangian Grassmannians. We also obtain new Laurent polynomial LandauGinzburg models for orthogonal Grassmannians, the Cayley plane and the Freudenthal variety.

P2GNet PoseGuided Point Cloud Generating Networks for 6DoF Object Pose Estimation ; Humans are able to perform fast and accurate object pose estimation even under severe occlusion by exploiting learned object model priors from everyday life. However, most recently proposed pose estimation algorithms neglect to utilize the information of object models, often end up with limited accuracy, and tend to fall short in cluttered scenes. In this paper, we present a novel learningbased model, underlinePoseGuided underlinePoint Cloud underlineGenerating Networks for 6D Object Pose Estimation P2GNet, designed to effectively exploit object model priors to facilitate 6D object pose estimation. We achieve this with an endtoend estimationbygeneration workflow that combines the appearance information from the RGBD image and the structure knowledge from object point cloud to enable accurate and robust pose estimation. Experiments on two commonly used benchmarks for 6D pose estimation, YCBVideo dataset and LineMOD dataset, demonstrate that P2GNet outperforms the stateoftheart method by a large margin and shows marked robustness towards heavy occlusion, while achieving realtime inference.

Electroweak Phase Transition in NonMinimal Higgs Sectors ; Higgs sector extensions beyond the Standard Model BSM provide additional sources of CP violation and further scalar states that help to trigger a strong first order electroweak phase transition SFOEWPT required to generate the observed baryon asymmetry of the Universe through electroweak baryogenesis. We investigate the CPviolating 2HiggsDoublet Model C2HDM and the NexttoMinimal 2HiggsDoublet Model N2HDM with respect to their potential to generate an SFOEWPT while being compatible with all relevant and recent theoretical and experimental constraints. The implications of an SFOEWPT on the collider phenomenology of the two models are analysed in detail in particular with respect to Higgs pair production. We provide benchmark points for parameter points that are compatible with an SFOEWPT and provide distinct diHiggs signatures.

Multipole Dark Energy ; While a scalar field with a pole in its kinetic term is often used to study the cosmological inflation, it can also play the role of dark energy, which is called the pole dark energy model. We propose a generalized model that the scalar field may have two or even multiple poles in the kinetic term and we call it the multipole dark energy. We find the poles can place some restrictions on the values of the original scalar field with noncanonical kinetic term. After transforming to the canonical form, we get a flat potential for the transformed new scalar field even if the original field has a steep one. It The latetime evolution of the universe is obtained explicitly for the two pole model, while dynamical analysis is performed for the multiple pole model. We find that it does have a stable attractor solution, which corresponds to the universe dominated by the potential of the scalar field.

Simultaneous Inference for Empirical Best Predictors with a Poverty Study in Small Areas ; Today, generalized linear mixed models are broadly used in many fields. However, the development of tools for performing simultaneous inference has been largely neglected in this domain. A framework for joint inference is indispensable to carry out statistically valid multiple comparisons of parameters of interest between all or several clusters. We therefore develop simultaneous confidence intervals and multiple testing procedures for empirical best predictors under generalized linear mixed models. In addition, we implement our methodology to study widely employed examples of mixed models, that is, the unitlevel binomial, the arealevel Poissongamma and the arealevel Poissonlognormal mixed models. The asymptotic results are accompanied by extensive simulations. A case study on predicting poverty rates illustrates applicability and advantages of our simultaneous inference tools.

RadionActivated Higgs Mechanism ; We study multiscalar models of radius stabilization, with an eye towards application to novel extradimensional models of symmetry breaking. With inspiration from holography, we construct a multiscalar effective potential that is a function of UVbrane values of the scalar fields, and that takes into account bulk gravitational backreaction. We study extrema of this potential, and additionally provide a superpotential method for generating static solutions for the extradimensional geometry. We apply these methods to some simple models of the Higgs mechanism where the Higgs itself plays a nontrivial role in radius stabilization. We conclude that mass mixing of the Higgs and radion is generic unless additional symmetries are imposed. We focus on models with moderate gap between the electroweak and KaluzaKlein scale, as required by phenomenological constraints. We note that tuning of the Higgs mass relative to the KK scale is related to various classes of tuning of 5D parameters, with different resulting spectra and phenomenologies.

Neural Shape Parsers for Constructive Solid Geometry ; Constructive Solid Geometry CSG is a geometric modeling technique that defines complex shapes by recursively applying boolean operations on primitives such as spheres and cylinders. We present CSGNe, a deep network architecture that takes as input a 2D or 3D shape and outputs a CSG program that models it. Parsing shapes into CSG programs is desirable as it yields a compact and interpretable generative model. However, the task is challenging since the space of primitives and their combinations can be prohibitively large. CSGNe uses a convolutional encoder and recurrent decoder based on deep networks to map shapes to modeling instructions in a feedforward manner and is significantly faster than bottomup approaches. We investigate two architectures for this task a vanilla encoder CNN decoder RNN and another architecture that augments the encoder with an explicit memory module based on the program execution stack. The stack augmentation improves the reconstruction quality of the generated shape and learning efficiency. Our approach is also more effective as a shape primitive detector compared to a stateoftheart object detector. Finally, we demonstrate CSGNet can be trained on novel datasets without program annotations through policy gradient techniques.

Axial Attention in Multidimensional Transformers ; We propose Axial Transformers, a selfattentionbased autoregressive model for images and other data organized as high dimensional tensors. Existing autoregressive models either suffer from excessively large computational resource requirements for high dimensional data, or make compromises in terms of distribution expressiveness or ease of implementation in order to decrease resource requirements. Our architecture, by contrast, maintains both full expressiveness over joint distributions over data and ease of implementation with standard deep learning frameworks, while requiring reasonable memory and computation and achieving stateoftheart results on standard generative modeling benchmarks. Our models are based on axial attention, a simple generalization of selfattention that naturally aligns with the multiple dimensions of the tensors in both the encoding and the decoding settings. Notably the proposed structure of the layers allows for the vast majority of the context to be computed in parallel during decoding without introducing any independence assumptions. This semiparallel structure goes a long way to making decoding from even a very large Axial Transformer broadly applicable. We demonstrate stateoftheart results for the Axial Transformer on the ImageNet32 and ImageNet64 image benchmarks as well as on the BAIR Robotic Pushing video benchmark. We open source the implementation of Axial Transformers.

Clustering and percolation on superpositions of Bernoulli random graphs ; A simple but powerful network model with n nodes and m partly overlapping layers is generated as an overlay of independent random graphs G1,dots,Gm with variable sizes and densities. The model is parameterised by a joint distribution Pn of layer sizes and densities. When m grows linearly and Pn to P as n to infty, the model generates sparse random graphs with a rich statistical structure, admitting a nonvanishing clustering coefficient together with a limiting degree distribution and clustering spectrum with tunable powerlaw exponents. Remarkably, the model admits parameter regimes in which bond percolation exhibits two phase transitions the first related to the emergence of a giant connected component, and the second to the appearance of gigantic singlelayer components.

A classical mechanical model of two interacting massless particles in de Sitter space and its quantization ; A conformally invariant model of two interacting massless particles in Minkowski space was proposed by Casalbuoni and Gomis 1. We generalize this model to the case of de Sitter space from the perspective of geodesic distance, in such a way that the resulting, generalized action reduces to the original action in a limit that de Sitter radius goes to infinity. We analyze the Hamiltonian formulation in accordance with Dirac's prescription for constrained Hamiltonian systems and carry out its subsequent canonical quantization in coordinate representation following DeWitt. As the result, we derive a fourthorder differential wave equation for bilocal fields that, in the infinite radius limit, reproduces one obtained in the original model for Minkowski space case.

Synthetic vascular structure generation for unsupervised pretraining in CTA segmentation tasks ; Large enough computed tomography CT data sets to train supervised deep models are often hard to come by. One contributing issue is the amount of manual labor that goes into creating ground truth labels, specially for volumetric data. In this research, we train a Unet architecture at a vessel segmentation task that can be used to provide insights when treating stroke patients. We create a computational model that generates synthetic vascular structures which can be blended into unlabeled CT scans of the head. This unsupervised approached to labelling is used to pretrain deep segmentation models, which are later finetuned on real examples to achieve an increase in accuracy compared to models trained exclusively on a handlabeled data set.

RBFFD analysis of 2D timedomain acoustic wave propagation in heterogeneous media ; Radial Basis Functiongenerated Finite Differences RBFFD is a popular variant of local strongform meshless methods that do not require a predefined connection between the nodes, making it easier to adapt nodedistribution to the problem under consideration. This paper investigates an RBFFD solution of timedomain acoustic wave propagation in the context of seismic modeling in the Earth's subsurface. Through a number of numerical tests, ranging from homogeneous to highlyheterogeneous velocity models including nonsmooth irregular topography, we demonstrate that the present approach can be further generalized to solve largescale seismic modeling and full waveform inversion problems in arbitrarily complex models enabling more robust interpretations of geophysical observations