Datasets:

sswt

/

arxiv_sample_50K

like 0

Using Cognitive Models to Train Warm Start Reinforcement Learning Agents for HumanComputer Interactions ; Reinforcement learning RL agents in humancomputer interactions applications require repeated user interactions before they can perform well. To address this cold start problem, we propose a novel approach of using cognitive models to pretrain RL agents before they are applied to real users. After briefly reviewing relevant cognitive models, we present our general methodological approach, followed by two case studies from our previous and ongoing projects. We hope this position paper stimulates conversations between RL, HCI, and cognitive science researchers in order to explore the full potential of the approach.

LightMBERT A Simple Yet Effective Method for Multilingual BERT Distillation ; The multilingual pretrained language models e.g, mBERT, XLM and XLMR have shown impressive performance on crosslingual natural language understanding tasks. However, these models are computationally intensive and difficult to be deployed on resourcerestricted devices. In this paper, we propose a simple yet effective distillation method LightMBERT for transferring the crosslingual generalization ability of the multilingual BERT to a small student model. The experiment results empirically demonstrate the efficiency and effectiveness of LightMBERT, which is significantly better than the baselines and performs comparable to the teacher mBERT.

Particle Creation in some LRS Bianchi I models ; In this work we consider particle creation by the expansion of the universe, using two Bianchi type I anisotropic models. The particles studied are of spin 0 and 12. The cosmological models have rotational symmetry, which allows us to solve exactly the equations of motion. The number density of the created particles is calculated with the method of Bogolubov transformations.

A posteriori error analysis of hybrid highorder method for the Stokes problem ; We present a residualbased a posteriori error estimator for the hybrid highorder HHO method for the Stokes model problem. Both the proposed HHO method and error estimator are valid in two and three dimensions and support arbitrary approximation orders on fairly general meshes. The upper bound and lower bound of the error estimator are proved, in which proof, the key ingredient is a novel stabilizer employed in the discrete scheme. By using the given estimator, adaptive algorithm of HHO method is designed to solve model problem. Finally, the expected theoretical results are numerically demonstrated on a variety of meshes for model problem.

Robustly Optimized and Distilled Training for Natural Language Understanding ; In this paper, we explore multitask learning MTL as a second pretraining step to learn enhanced universal language representation for transformer language models. We use the MTL enhanced representation across several natural language understanding tasks to improve performance and generalization. Moreover, we incorporate knowledge distillation KD in MTL to further boost performance and devise a KD variant that learns effectively from multiple teachers. By combining MTL and KD, we propose Robustly Optimized and Distilled ROaD modeling framework. We use ROaD together with the ELECTRA model to obtain stateoftheart results for machine reading comprehension and natural language inference.

GumbelAttention for Multimodal Machine Translation ; Multimodal machine translation MMT improves translation quality by introducing visual information. However, the existing MMT model ignores the problem that the image will bring information irrelevant to the text, causing much noise to the model and affecting the translation quality. This paper proposes a novel GumbelAttention for multimodal machine translation, which selects the textrelated parts of the image features. Specifically, different from the previous attentionbased method, we first use a differentiable method to select the image information and automatically remove the useless parts of the image features. Experiments prove that our method retains the image features related to the text, and the remaining parts help the MMT model generates better translations.

Exactly solvable models for U1 symmetryenriched topological phases ; We propose a general construction of commuting projector lattice models for 2D and 3D topological phases enriched by U1 symmetry, with finitedimensional Hilbert space per site. The construction starts from a commuting projector model of the topological phase and decorates U1 charges to the state space in a consistent manner. We show that all 2D U1 symmetryenriched topological phases which allow gapped boundary without breaking symmetry, can be realized through our construction. We also construct a large class of 3D topological phases with U1 symmetry fractionalized on particles or loop excitations.

Applications of Controlled Sweeping Processes to Nonlinear Crowd Motion Models with Obstacles ; This paper mainly focuses on solving the dynamic optimization of the planar controlled crowd motion models with obstacles which is an application of a class of optimal control problems governed by a general perturbed nonconvex sweeping process. This can be considered as a significant extension of the previous work regarding the controlled crowd motion models, where the obstacles have not been considered. The necessary optimality conditions for the problem under consideration are established and illustrated by a nontrivial example of practical importance.

Wealth distribution in modern societies collected data and a master equation approach ; A meanfield like stochastic evolution equation with growth and reset terms LGGR model is used to model wealth distribution in modern societies. The stationary solution of the model leads to an analytical form for the density function that is successful in describing the observed data for all wealth categories. In the limit of high wealth values the proposed density function has the accepted TsallisPareto shape. Our results are in agreement with the predictions of an earlier approach based on a meanfield like wealth exchange process.

Building a Swedish OpenDomain Conversational Language Model ; We present ongoing work of evaluating the, to our knowledge, first large generative language model trained to converse in Swedish, using data from the online discussion forum Flashback. We conduct a human evaluation pilot study that indicates the model is often able to respond to conversations in both a humanlike and informative manner, on a diverse set of topics. While data from online forums can be useful to build conversational systems, we reflect on the negative consequences that incautious application might have, and the need for taking active measures to safeguard against them.

Light and thermodynamics the threelevel laser as an endoreversible heat engine ; In the past, a number of heat engine models have been devised to apply the principles of thermodynamics to a laser. The best one known is the model using a negative temperature to describe population inversion. In this paper, we present a new temperature scale not based on reservoir temperatures. This is realized by revealing a formal mathematical similarity between the expressions for the optimum power generated by an endoreversible heat engine and the optimum outputcoupled power from a threelevel laser resonator. As a consequence, the theory of endoreversibility can be applied to enable a new efficiency analysis of the cooling of highpower lasers. We extend the endoreversibility concept also to the fourlevel laser model.

Multipole extension for elliptic models of interacting integrable tops ; We review and give detailed description for rm glNM Gaudin models related to holomorphic vector bundles of rank NM and degree N over elliptic curve with n punctures. Then we introduce their generalizations constructed by means of Rmatrices satisfying the associative YangBaxter equation. A natural extension of the obtained models to the Schlesinger systems is given as well.

MOROCCO Model Resource Comparison Framework ; The new generation of pretrained NLP models push the SOTA to the new limits, but at the cost of computational resources, to the point that their use in real production environments is often prohibitively expensive. We tackle this problem by evaluating not only the standard quality metrics on downstream tasks but also the memory footprint and inference time. We present MOROCCO, a framework to compare language models compatible with textttjiant environment which supports over 50 NLU tasks, including SuperGLUE benchmark and multiple probing suites. We demonstrate its applicability for two GLUElike suites in different languages.

Entropy production estimate for the ESBGK model with the correct Prandtl number ; In this paper, we establish the entropyentropy production estimate for the ESBGK model, a generalized version of the BGK model of the Boltzmann equation introduced for better approximation in the NavierStokes limit. Our result improves the previous entropy production estimate 39 in that 1 the full range of Prandtl parameters 12leqnu 1 including the critical case nu12 is covered, and 2 a sharper entropy production bound is obtained. An explicit characterization of the coefficient of the entropyentropy production estimate is also presented.

Quark model analysis of the Weinberg operator contribution to the nucleon EDM ; The Weinberg operator chromoelectric dipole moment of gluon is a CP violating quantity generated in many candidates of new physics beyond the standard model, and it contributes to observables such as the electric dipole moments EDM of the neutron or atoms which are currently measured in experiments. In this proceedings contribution, we report on our result of the evaluation of the Weinberg operator contribution to the nucleon EDM in the nonrelativistic quark model using the Gaussian expansion method.

Emergent Supersymmetry on the Edges ; The WZW models describe the dynamics of the edge modes of ChernSimons theories in three dimensions. We explore the WZW models which can be mapped to supersymmetric theories via the generalized JordanWigner transformation. Some of such models have supersymmetric Ramond vacua, but the others break the supersymmetry spontaneously. We also make a comment on recent proposals that the ReadRezayi states at filling fraction nu12,23 are able to support supersymmetry.

Deposition control of model glasses with surfacemediated orientational order ; We introduce a minimal model of solidforming anisotropic molecules that displays, in thermal equilibrium, surface orientational order without bulk orientational order. The model reproduces the nonequilibrium behavior of recent experiments in that a bulk nonequilibrium structure grown by deposition contains regions of orientational order characteristic of the surface equilibrium. This order is deposited in general in a nonuniform way, because of the emergence of a growthpoisoning mechanism that causes equilibrated surfaces to grow slower than nonequilibrated surfaces. We use evolutionary methods to design oscillatory protocols able to grow nonequilibrium structures with uniform order, demonstrating the potential of protocol design for the fabrication of this class of materials.

Hamiltonian analysis of the Schroedinger field coupled with dynamic nonrelativistic gravity ; Using the recently mooted Galilean gauge theory we have constructed the model for the Schroedinger field interacting wuth gravity which is also dynamical. The dynamics of gravity is dictated by the Newtonian action in the NewtonCartan spacetime. The theory is highly constrained . An elaborate analysis of the constraints of the theory have been performed. The symmetries are explicitly verified and the uniqueness of the model has been established. To the best of our knowledge both the model and its constraints structure are unique in the literature

The Minimal Supersymmetric Universal Seesaw Mechanism MSUSM ; We build a supersymmetric model with SU2Lotimes SU2Rotimes U1BL electroweak gauge symmetry, where SU2L is the lefthanded currents while SU2R is the righthanded currents and B and L are the usual baryonic and leptonic numbers. We can generate an universal seesaw mechanism to get masses for all the usual fermions in this model, it means quarks and leptons, and also explain the mixing experimental data. We will also to study the masses of the Gauge Bosons and also the masses of all usual scalars of this model.

Risk Model of German Corona Warning App Reloaded ; In this paper we discuss the risk model of the German Corona Warning App CWA in two versions. Both are based on a general semiquantitative risk approach that is not state of the art anymore and for some application domains even deprecated. However, it turns out that the CWA uses a much more limited model, that does not even assess risk, but relies only on one parameter, a weighted exposure time. It is shown that the CWA grossly underestimates even this parameter and so may reassure the users wrongly. As the CWA also has other systematic limitations and shortcomings it is advised not to rely on its results but rather on Covid testing and vaccination.

Phase Transitions in Ehrenfest Urns Model with Interactions Coexistence of uniform and nonuniform states ; A model based on the classic noninteracting Ehrenfest urn model with twourns is generalized to M urns with the introduction of interactions for particles within the same urn. As the interparticle interaction strength is varied, phases of different levels of nonuniformity emerge and their stabilities are calculated analytically. In particular, coexistence of locally stable uniform and nonuniform phases connected by firstorder transition occurs. The phase transition threshold and energy barrier can be derived exactly together with the phase diagram obtained analytically. These analytic results are further confirmed by Monte Carlo simulations.

HighFrequency aware Perceptual Image Enhancement ; In this paper, we introduce a novel deep neural network suitable for multiscale analysis and propose efficient modelagnostic methods that help the network extract information from highfrequency domains to reconstruct clearer images. Our model can be applied to multiscale image enhancement problems including denoising, deblurring and single image superresolution. Experiments on SIDD, Flickr2K, DIV2K, and REDS datasets show that our method achieves stateoftheart performance on each task. Furthermore, we show that our model can overcome the oversmoothing problem commonly observed in existing PSNRoriented methods and generate more natural highresolution images by applying adversarial training.

Gender Bias Amplification During SpeedQuality Optimization in Neural Machine Translation ; Is bias amplified when neural machine translation NMT models are optimized for speed and evaluated on generic test sets using BLEU We investigate architectures and techniques commonly used to speed up decoding in Transformerbased models, such as greedy search, quantization, average attention networks AANs and shallow decoder models and show their effect on gendered noun translation. We construct a new gender bias test set, SimpleGEN, based on gendered noun phrases in which there is a single, unambiguous, correct answer. While we find minimal overall BLEU degradation as we apply speed optimizations, we observe that gendered noun translation performance degrades at a much faster rate.

Improving Formality Style Transfer with ContextAware Rule Injection ; Models pretrained on largescale regular text corpora often do not work well for usergenerated data where the language styles differ significantly from the mainstream text. Here we present ContextAware Rule Injection CARI, an innovative method for formality style transfer FST. CARI injects multiple rules into an endtoend BERTbased encoder and decoder model. It learns to select optimal rules based on context. The intrinsic evaluation showed that CARI achieved the new highest performance on the FST benchmark dataset. Our extrinsic evaluation showed that CARI can greatly improve the regular pretrained models' performance on several tweet sentiment analysis tasks.

Existence of replicasymmetry breaking in quantum glasses ; By controlling quantum fluctuations via the FalkBruch inequality we give the first rigorous argument for the existence of a spinglass phase in the quantum SherringtonKirkpatrick model with a transverse magnetic field if the temperature and the field are sufficiently low. The argument also applies to the generalization of the model with multispin interactions, sometimes dubbed as transverse pspin model.

Can Generative Pretrained Language Models Serve as Knowledge Bases for Closedbook QA ; Recent work has investigated the interesting question using pretrained language models PLMs as knowledge bases for answering open questions. However, existing work is limited in using small benchmarks with high testtrain overlaps. We construct a new dataset of closedbook QA using SQuAD, and investigate the performance of BART. Experiments show that it is challenging for BART to remember training facts in high precision, and also challenging to answer closedbook questions even if relevant knowledge is retained. Some promising directions are found, including decoupling the knowledge memorizing process and the QA finetune process, forcing the model to recall relevant knowledge when question answering.

Improving Unsupervised Dialogue Topic Segmentation with UtterancePair Coherence Scoring ; Dialogue topic segmentation is critical in several dialogue modeling problems. However, popular unsupervised approaches only exploit surface features in assessing topical coherence among utterances. In this work, we address this limitation by leveraging supervisory signals from the utterancepair coherence scoring task. First, we present a simple yet effective strategy to generate a training corpus for utterancepair coherence scoring. Then, we train a BERTbased neural utterancepair coherence model with the obtained training corpus. Finally, such model is used to measure the topical relevance between utterances, acting as the basis of the segmentation inference. Experiments on three public datasets in English and Chinese demonstrate that our proposal outperforms the stateoftheart baselines.

Parametrization of renormalized models for singular stochastic PDEs ; Let mathscrT be the regularity structure associated with a given system of singular stochastic PDEs. The paracontrolled representation of the sf Pi map provides a linear parametrization of the nonlinear space of admissible models sf Mg,Pi on mathscrT, in terms of the family of pararemainders used in the representation. We give an explicit description of the action of the most general class of renormalization schemes presently available on the parametrization space of the space of admissible models. The action is particularly simple for renormalization schemes associated with degree preserving preparation maps; the BHZ renormalization scheme has that property.

On Graphical Models and Convex Geometry ; We introduce a mixturemodel of beta distributions to identify significant correlations among P predictors when P is large. The method relies on theorems in convex geometry, which we use to show how to control the error rate of edge detection in graphical models. Our betaMix' method does not require any assumptions about the network structure, nor does it assume that the network is sparse. The results in this article hold for a wide class of data generating distributions that include lighttailed and heavytailed spherically symmetric distributions.

Complexitybased partitioning of CSFI problem instances with Transformers ; In this paper, we propose a twosteps approach to partition instances of the Conjunctive Normal Form CNF Syntactic Formula Isomorphism problem CSFI into groups of different complexity. First, we build a model, based on the Transformer architecture, that attempts to solve instances of the CSFI problem. Then, we leverage the errors of such model and train a second Transformerbased model to partition the problem instances into groups of different complexity, thus detecting the ones that can be solved without using too expensive resources. We evaluate the proposed approach on a pseudorandomly generated dataset and obtain promising results. Finally, we discuss the possibility of extending this approach to other problems based on the same type of textual representation.

Extracting the parton distribution functions evolution equations using the stochastic modeling in the nonequilibrium statistical mechanics ; In this paper, using the stochastic modeling of the nonequilibrium statistical mechanics in the momentum space, the evolution equations of the parton distribution functions PDF usually used in the hadrons phenomenology are generated. These stochastic modeling PDF evolution equations are the same as those of the DokshitzerGribovLipatovAltarelliParisi DGLAP ones, but they can be obtained by a more simplistic mathematical procedure based on the nonequilibrium statistical mechanics and the theory of Markov processes.

Reducedform framework for multiple ordered default times under model uncertainty ; In this paper we introduce a sublinear conditional operator with respect to a family of possibly nondominated probability measures in presence of multiple ordered default times. In this way we generalize the results of 5, where a reducedform framework under model uncertainty for a single default time is developed. Moreover, we use this operator for the valuation of credit portfolio derivatives under model uncertainty.

An Energybased, always Index le1 and Structurally Amenable Electrical Circuit Model ; Combining three themes portHamiltonian energybased modelling, structural analysis as used in the circuit world, and structural analysis of general differentialalgebraic equations, we form a new model for electrical circuits, the compact portHamiltonian equations. They have remarkable simplicity and symmetry, and always have index at most 1 and other good numerical properties. The method has been implemented in Matlab. We give proofs and numerical results.

Hawking radiation and black hole gravitational back reaction A quantum geometrodynamical simplified model ; The purpose of this paper is to analyse the back reaction problem, between Hawking radiation and the black hole, in a simplified model for the black hole evaporation in the quantum geometrodynamics context. The idea is to transcribe the most important characteristics of the WheelerDeWitt equation into a Schrodinger's type of equation. Subsequently, we consider Hawking radiation and black hole quantum states evolution under the influence of a potential that includes back reaction. Finally, entropy is estimated as a measure of the entanglement between the black hole and Hawking radiation states in this model.

A Game Interface to Study Semantic Grounding in TextBased Models ; Can language models learn grounded representations from text distribution alone This question is both central and recurrent in natural language processing; authors generally agree that grounding requires more than textual distribution. We propose to experimentally test this claim if any two words have different meanings and yet cannot be distinguished from distribution alone, then grounding is out of the reach of textbased models. To that end, we present early work on an online game for the collection of human judgments on the distributional similarity of word pairs in five languages. We further report early results of our data collection campaign.

Regularizing Transformers With Deep Probabilistic Layers ; Language models LM have grown with nonstop in the last decade, from sequencetosequence architectures to the stateoftheart and utter attentionbased Transformers. In this work, we demonstrate how the inclusion of deep generative models within BERT can bring more versatile models, able to impute missingnoisy words with richer text or even improve BLEU score. More precisely, we use a Gaussian Mixture Variational Autoencoder GMVAE as a regularizer layer and prove its effectiveness not only in Transformers but also in the most relevant encoderdecoder based LM, seq2seq with and without attention.

Stationarity and inference in multistate promoter models of stochastic gene expression via stickbreaking measures ; In a general stochastic multistate promoter model of dynamic mRNAprotein interactions, we identify the stationary joint distribution of the promoter state, mRNA, and protein levels through an explicit stickbreaking' construction of interest in itself. This derivation is a constructive advance over previous work where the stationary distribution is solved only in restricted cases. Moreover, the stickbreaking construction allows to sample directly from the stationary distribution, permitting inference procedures and model selection. In this context, we discuss numerical Bayesian experiments to illustrate the results.

Mimicking the CDM Universe through inhomogeneous spacetime ; Starting from an inhomogeneous spacetime model of the universe we could recreate a scenario of recent time accelerating universe dominated by Dark Energy type of fluid. The background matter component of such a universe was considered to be made up of a combination of an anisotropic fluid, a barotropic fluid and the presureless cold dark matter. It was found that inhomogeneity exhibits itself as the curvature term in such a universe. We corroborated our model with recent supernova IaJLA data together with H0 data and BAO data. Cosmographic analysis of the dynamical variables further show that the model can mimic the LambdaCDM cosmology very closely.

On the Limits of Minimal Pairs in Contrastive Evaluation ; Minimal sentence pairs are frequently used to analyze the behavior of language models. It is often assumed that model behavior on contrastive pairs is predictive of model behavior at large. We argue that two conditions are necessary for this assumption to hold First, a tested hypothesis should be wellmotivated, since experiments show that contrastive evaluation can lead to false positives. Secondly, test data should be chosen such as to minimize distributional discrepancy between evaluation time and deployment time. For a good approximation of deploymenttime decoding, we recommend that minimal pairs are created based on machinegenerated text, as opposed to humanwritten references. We present a contrastive evaluation suite for EnglishGerman MT that implements this recommendation.

A domainspecific modeling and analysis environment for complex IoT applications ; To cope with the complexities found in the Internet of Things domain, designers and developers of IoT applications demand practical tools. Several modeldriven engineering methodologies and tools have been developed to address such difficulties, but few of them address the analysis aspects. In this extended abstract, we introduce CHESSIoT, a domainspecific modeling environment for complex IoT applications. In addition, the existing supported realtime analysis mechanism, as well as a proposed code generation approach, are presented

Asymptotics for multifactor Volterra type stochastic volatility models ; We study multidimensional stochastic volatility models in which the volatility process is a positive continuous function of a continuous multidimensional Volterra process that can be not selfsimilar. The main results obtained in this paper are a generalization of the results due, in the onedimensional case, to Cellupica and Pacchiarotti M. Cellupica and B. Pacchiarotti 2021 Pathwise Asymptotics for Volterra Type Stochastic Volatility Models. Journal of Theoretical Probability, 342682727. We state some pathwise and finitedimensional large deviation principles for the scaled logprice and as a consequence some pathwise and finitedimensional shorttime large deviation principles.

Effective potential of scalartensor gravity with quartic selfinteraction of scalar field ; Oneloop effective potential of scalartensor gravity with a quartic scalar field selfinteraction is evaluated up to first postMinkowskian order. The potential develops an instability in the strong field regime which is expected from an effective theory. Depending on model parameters the instability region can be exponentially far in a strong field region. Possible applications of the model for inflationary scenarios are highlighted. It is shown that the model can enter the slowroll regime with a certain set of parameters.

Global wellposedness for the nonlinear generalized parabolic Anderson model equation ; We study the global existence of the singular nonlinear parabolic Anderson model equation on 2dimensional tours mathbbT2. The method is based on paracontrolled distribution and renormalization. After split the original nonlinear parabolic Anderson model equation into two simple equations, we prove the global wellposedness by some a priori estimates and smooth approximations. Furthermore, we prove the uniqueness of the solution by using classical energy estimates.

Interactive Probing of Multivariate Time Series Prediction Models A Case of Freight Rate Analysis ; We present an interactive probing tool to create, modify and analyze whatif scenarios for multivariate time series models. The solution is applied to freight trading, where analysts can carry out sensitivity analysis on freight rates by changing demand and supplyrelated econometric variables and observing their resultant effects on freight indexes. We utilize various visualization techniques to enable intuitive scenario creation, alteration, and comprehension of time series inputs and model predictions. Our tool proved to be useful to the industry practitioners, demonstrated by a case study where freight traders are given hypothetical market scenarios and successfully generated quantitative freight index projection with confidence.

Deep Social Force ; The Social Force model introduced by Helbing and Molnar in 1995 is a cornerstone of pedestrian simulation. This paper introduces a differentiable simulation of the Social Force model where the assumptions on the shapes of interaction potentials are relaxed with the use of universal function approximators in the form of neural networks. Classical forcebased pedestrian simulations suffer from unnatural locking behavior on headon collision paths. In addition, they cannot model the bias of pedestrians to avoid each other on the right or left depending on the geographic region. My experiments with more general interaction potentials show that potentials with a sharp tip in the front avoid locking. In addition, asymmetric interaction potentials lead to a left or right bias when pedestrians avoid each other.

What should be the ontology for the Standard Model ; Although the Standard Model of particle physics is usually formulated in terms of fields, it can be equivalently formulated in terms of particles and strings. In this picture particles and open strings are always coupled. This offers an intuitive and graphical explanation for the otherwise mysterious gauge symmetry. In addition, the particlestring formulation avoids introducing redundant path integral configurations that are present in the field formulation. For its explanatory power and economy, the particlestring ontology may be preferred over the field ontology for the Standard Model.

Wavelet Estimation for Factor Models with TimeVarying Loadings ; We introduce a highdimensional factor model with timevarying loadings. We cover both stationary and nonstationary factors to increase the possibilities of applications. We propose an estimation procedure based on two stages. First, we estimate common factors by principal components. In the second step, considering the estimated factors as observed, the timevarying loadings are estimated by an iterative generalized least squares procedure using wavelet functions. We investigate the finite sample features by some Monte Carlo simulations. Finally, we apply the model to study the Nord Pool power market's electricity prices and loads.

Matrix models and nonAbelian T dual of AdS5 times S5 ; We compute Euclidean Wilson loops for BMN Plane Wave Matrix Models. The stringy counterpart of these Wilson loops corresponds to various semiclassical open string embedding that probe a class of half BPS geometries in Type IIA supergravity. These geometries fall under the general category of LinLunin and Maldacena LLM. As a special class of solutions within the LLM category, we construct Wilson operators corresponding to the nonAbelian T dual NATD of AdS5 times S5 . Our analysis shows close agreement between matrix model and string theory calculations.

NonGaussianity in DHOST inflation ; DHOST inflation models where deviations from a pure de Sitter background are induced by an axionlike potential can lead to large nonGaussianities. We investigate the nature of nonGaussianities in these models and compare to the results given by the Planck experiment. The overlap between the DHOST nonGaussianities and the equilateral, orthogonal and local templates can be rendered arbitrarily small. On the other hand, this does not preclude DHOST models from showing large nonGaussianities as exemplified by their reduced bispectrum. As a result, they could be probed by future experiments and also by a more thorough analysis of the existing Planck data.

A Categorical Semantics of Fuzzy Concepts in Conceptual Spaces ; We define a symmetric monoidal category modelling fuzzy concepts and fuzzy conceptual reasoning within Gardenfors' framework of conceptual convex spaces. We propose logconcave functions as models of fuzzy concepts, showing that these are the most general choice satisfying a criterion due to Gardenfors and which are wellbehaved compositionally. We then generalise these to define the category of logconcave probabilistic channels between convex spaces, which allows one to model fuzzy reasoning with noisy inputs, and provides a novel example of a Markov category.

DistributionFree Bayesian multivariate predictive inference ; We introduce a comprehensive Bayesian multivariate predictive inference framework. The basis for our framework is a hierarchical Bayesian model, that is a mixture of finite Polya trees corresponding to multiple dyadic partitions of the unit cube. Given a sample of observations from an unknown multivariate distribution, the posterior predictive distribution is used to model and generate future observations from the unknown distribution. We illustrate the implementation of our methodology and study its performance on simulated examples. We introduce an algorithm for constructing conformal prediction sets, that provide finite sample probability assurances for future observations, with our Bayesian model.

A Lattice Model for Super LLT Polynomials ; We introduce a solvable lattice model for supersymmetric LLT polynomials, also known as super LLT polynomials, based upon particle interactions in super nribbon tableaux. Using operators on a Fock space, we prove a Cauchy identity for super LLT polynomials, simultaneously generalizing the Cauchy and dual Cauchy identities for LLT polynomials. Lastly, we construct a solvable semiinfinite Cauchy lattice model with a surprising YangBaxter equation and examine its connections to the Cauchy identity.

Fractality in Cosmic Topology Models with Spectral Action Gravity ; We consider cosmological models based on the spectral action formulation of modified gravity. We analyze the coupled effects, in this model, of the presence of nontrivial cosmic topology and of fractality in the large scale structure of spacetime. We show that the topology constrains the possible fractal structures, and in turn the correction terms to the spectral action due to fractality distinguish the various cosmic topology candidates, with effects detectable in a slowroll inflation scenario, through the power spectra of the scalar and tensor fluctuations. We also discuss explicit effects of the presence of fractal structures on the gravitational waves equations.

Nonlinear Reduced DNN Models for State Estimation ; We propose in this paper a data driven state estimation scheme for generating nonlinear reduced models for parametric families of PDEs, directly providing datatostate maps, represented in terms of Deep Neural Networks. A major constituent is a sensorinduced decomposition of a modelcompliant Hilbert space warranting approximation in problem relevant metrics. It plays a similar role as in a Parametric Background Data Weak framework for state estimators based on Reduced Basis concepts. Extensive numerical tests shed light on several optimization strategies that are to improve robustness and performance of such estimators.

Scalable knowledge base completion with superposition memories ; We present Harmonic Memory Networks HMem, a neural architecture for knowledge base completion that models entities as weighted sums of pairwise bindings between an entity's neighbors and corresponding relations. Since entities are modeled as aggregated neighborhoods, representations of unseen entities can be generated on the fly. We demonstrate this with two new datasets WNGen and FBGen. Experiments show that the model is SOTA on benchmarks, and flexible enough to evolve without retraining as the knowledge graph grows.

Ultra Light OCR Competition Technical Report ; Ultra Light OCR Competition is a Chinese scene text recognition competition jointly organized by CSIG China Society of Image and Graphics and Baidu, Inc. In addition to focusing on common problems in Chinese scene text recognition, such as long text length and massive characters, we need to balance the tradeoff of model scale and accuracy since the model size limitation in the competition is 10M. From experiments in aspects of data, model, training, etc, we proposed a general and effective method for Chinese scene text recognition, which got us second place among over 100 teams with accuracy 0.817 in TestB dataset. The code is available at httpsaistudio.baidu.comaistudioprojectdetail2159102.

Model based Multiagent Reinforcement Learning with Tensor Decompositions ; A challenge in multiagent reinforcement learning is to be able to generalize over intractable stateaction spaces. Inspired from Tesseract Mahajan et al., 2021, this position paper investigates generalisation in stateaction space over unexplored stateaction pairs by modelling the transition and reward functions as tensors of low CPrank. Initial experiments on synthetic MDPs show that using tensor decompositions in a modelbased reinforcement learning algorithm can lead to much faster convergence if the true transition and reward functions are indeed of low rank.

The nonlinear anisotropic model of the Universe with the linear potential ; In the Bianchi I cosmology some subclasses of the Horndeski theory allow for the nonstandard anisotropy behavior. For example, the anisotropy is damped near the initial singularity instead of tending to infinity. In this article, we analyze the nonlinear anisotropic model with the linear potential.We have considered an example of such theory, for which the anisotropy is always finite. The anisotropy reaches its a maximum value at the initial moment. The anisotropy suppression occurs during the inflationary stage, and it approaches zero at later times. This cosmological model does not contain the singular point.

Differential models for the Anderson dual to bordism theories and invertible QFT's, II ; This is the second part of the work on differential models of the Anderson duals to the stable tangential Gbordism theories IOmegaG, motivated by classifications of invertible QFT's. Using the model constructed in the first part citeYamashitaYonekura2021, in this paper we show that pushforwards in generalized differential cohomology theories induces transformations between differential cohomology theories which refine the Anderson duals to multiplicative genera. This gives us a unified understanding of an important class of elements in the Anderson duals with physical origins.

Algebraic algorithm for direct sampling from toric models ; We show that Pfaffians or contiguity relations of hypergeometric functions of several variables give a direct sampling algorithm from toric models in statistics, which is a Markov chain on a lattice generated by a matrix A. A correspondence among graphical toric models and Ahypergeometric system is discussed and we give a sum formula of special values of Ahypergeometric polynomials. Some hypergeometric series which are interesting in view of statistics are presented.

Recent jet and jet substructure measurements at the LHC, and ML based tagging ; Recent jet and jet substructure measurements at the LHC, and of machinelearningbased tagging techniques are presented using protonproton collision data collected by the ATLAS and CMS experiments at CERN's Large Hadron Collider. These measurements are crucial for precise tests of electroweak and pQCD calculations and searches for physics beyond the Standard Model. The measurements are compared with several Monte Carlo event generator predictions which provide valuable input to the tuning of perturbative and nonperturbative models and to constraining model parameters of advanced partonshower Monte Carlo programs.

New Physics at a multiTeV Collider ; We present the results of the physics reach of a multiTeV muon collider for popular models of physics beyond the Standard Model. We include also details about a model predicting the scalar dark matter in the spectrum. Finally we present some preliminary results about the Effective Vector Approximation and its implementation in the Monte Carlo generator sc MadGraph5aMCNLO.

Syzygies of string modules for special biserial algebras ; We present discrete models of special biserial SB algebras and their string modules, drawing inspiration from cellular automata, and cast new light on patterns among syzygies. We explore applications of our models to open questions in homological algebra regarding certain triangulated subcategories of derived categories, with implications for the finitistic dimension conjectures. More pertinently, our models provide the inner workings for a new, original GAP package called SBStrips, written and implemented by the author. Its source code is freely available online and its documentation is included as an appendix. The package calculates syzygies of string modules and much more besides using specialised methods much more efficient than the general methods currently employed by the QPA package.

Finite thermostats in classical and quantum nonequilibrium ; Abstract Models for studying systems in stationary states but out of equilibrium have often empirical nature and very often break the fundamental time reversal symmetry. Here a formal interpretation will be discussed of the widespread idea that, in any event, the particular friction model choice should not matter physically. The proposal is, quite generally, that for the same physical system a time reversible model should be possible. Examples about the NavierStokes equations are given.

On the derived models of self iterable universes ; We show that if the universe is selfiterable and kappa is an inaccessible limit of Woodin cardinal then ADR Theta is regular holds in the derived model at kappa. The proof is finestructure free, and only assumes basic knowledge of iteration trees and iteration strategies. Our proof can be viewed as the finestructure free version of the wellknown fact that ADR Theta is regular is true in the derived models of hod mice that have inaccessible limit of Woodin cardinals see for example 6. However, the proof uses a different set of ideas and is more general.

Adversarial attacks on voter model dynamics in complex networks ; This study investigates adversarial attacks conducted to distort voter model dynamics in complex networks. Specifically, a simple adversarial attack method is proposed to hold the state of opinions of an individual closer to the target state in the voter model dynamics. This indicates that even when one opinion is the majority, the vote outcome can be inverted i.e., the outcome can lean toward the other opinion by adding extremely small hardtodetect perturbations strategically generated in social networks. Adversarial attacks are relatively more effective in complex large and dense networks. These results indicate that opinion dynamics can be unknowingly distorted.

The voter model with a slow membrane ; We introduce the voter model on the infinite lattice with a slow membrane and investigate its hydrodynamic behavior and nonequilibrium fluctuations. The model is defined as follows a voter adopts one of its neighbors' opinion at rate one except for neighbors crossing the hyperplane xx1 12, where the rate is alpha Nbeta. Above, alpha0,,beta geq 0 are two parameters and N is the scaling parameter. The hydrodynamic equation turns out to be heat equation with various boundary conditions depending on the value of beta. For the nonequilibrium fluctuations, the limit is described by generalized OrnsteinUhlenbeck process with certain boundary condition corresponding to the hydrodynamic equation.

Cosmological Particle Production A Review ; This article will review quantum particle creation in expanding universes. The emphasis will be on the basic physical principles and on selected applications to cosmological models. The needed formalism of quantum field theory in curved spacetime will be summarized, and applied to the example of scalar particle creation in a spatially flat universe. Estimates for the creation rate will be given and applied to inflationary cosmology models. Analog models which illustrate the same physical principles and may be experimentally realizable are also discussed.

Cosmological model with time varying deceleration parameter in FR,G gravity ; In this paper, we study the dynamical behaviour of the Universe in the FR,G theory of gravity, where R and G respectively denote the Ricci scalar and GaussBonnet invariant. Our wide analysis encompasses the energy conditions, cosmographic parameters, Omz diagnostic, stability and the viability of reconstructing the referred model through a scalar field formalism. The model obtained here shows the quintessence like behaviour at late times.

Simulation of Gaussian random field in a ball ; The presented paper is devoted to statistical modeling of Gaussian scalar real random fields inside a threedimensional sphere ball. We propose a statistical model describing the spatial heterogeneity in a unit ball and a numerical procedure for generating an ensemble of corresponding random realizations. The accuracy of the presented approach is corroborated by the numerical comparison of the estimated and analytical covariance functions. Our approach is flexible with respect to the assumed radial and angular covariance function. We illustrate the effect of the covariance model parameters based on numerical examples of random field realizations. The presented statistical simulation technique can be applied, for example, to the inference of the 3D spatial heterogeneity in the Earth and other planets.

U1 gauged boson stars in the EinsteinFriedbergLeeSirlin model ; We consider spherically symmetric U1 gauged boson stars in the twocomponent scalar FriedbergLeeSirlin model with a symmetry breaking potential in 31 dimensional spacetime. Depending on the relative strength of gravity and the electromagnetic interaction, the resulting boson stars exhibit either the typical properties of ungauged boson stars, or their behavior resembles the pattern found for gauged Qballs of the FriedbergLeeSirlin model in flat spacetime, both for a finite and a vanishing potential.

Applying SoftTriple Loss for Supervised Language Model Fine Tuning ; We introduce a new loss function TripleEntropy, to improve classification performance for finetuning general knowledge pretrained language models based on crossentropy and SoftTriple loss. This loss function can improve the robust RoBERTa baseline model finetuned with crossentropy loss by about 0.02 2.29. Thorough tests on popular datasets indicate a steady gain. The fewer samples in the training dataset, the higher gain thus, for smallsized dataset it is 0.78, for mediumsized 0.86 for large 0.20 and for extralarge 0.04.

Pattern capacity of a single quantum perceptron ; Recent developments in Quantum Machine Learning have seen the introduction of several models to generalize the classical perceptron to the quantum regime. The capabilities of these quantum models need to be determined precisely in order to establish if a quantum advantage is achievable. Here we use a statistical physics approach to compute the pattern capacity of a particular model of quantum perceptron realized by means of a continuous variable quantum system.

Stability analysis of an eight parameter SIRtype model including loss of immunity, and disease and vaccination fatalities ; We revisit here a landmark five parameter SIRtype model of DvdD93, Sec. 4, which is maybe the simplest example where a complete picture of all cases, including nontrivial bistability behavior, may be obtained using simple tools. We also generalize it by adding essential vaccination and vaccinationinduced death parameters, with the aim of revealing the role of vaccination and its possible failure. The main result is Theorem 5, which describes the stability behavior of our model in all possible cases.

The 1jet determination of stationary discs attached to generic CR submanifolds ; The existence of a nondefective stationary disc attached to a nondegenerate model quadric in CN is a necessary condition to ensure the unique 1jet determination of the lifts of a key family of stationary discs. In this paper, we give an elementary proof of the equivalence when the model quadric is strongly pseudoconvex, recovering a result of Tumanov. Our proof is based on the explicit expression of stationary discs, and opens up a conjecture for the unique 1jet determination to hold when the model is not necessarily strongly pseudoconvex.

Wellposedness and singularity formation for the Kolmogorov twoequation model of turbulence in 1D ; We study the Kolomogorov twoequation model of turbulence in one space dimension. Two are the main results of the paper. First of all, we establish a local wellposedness theory in Sobolev spaces even in the case of vanishing mean turbulent kinetic energy. Then, we show that, in general, those solutions must blow up in finite time. To the best of our knowledge, these results are the first establishing the wellposedness of the system for vanishing initial data and the occurence of finite time singularities for the model under study.

Rough multifactor volatility for SPX and VIX options ; We provide explicit smalltime formulae for the atthemoney implied volatility, skew and curvature in a large class of models, including rough volatility models and their multifactor versions. Our general setup encompasses both European options on a stock and VIX options, thereby providing new insights on their joint calibration. The tools used are essentially based on Malliavin calculus for Gaussian processes. We develop a detailed theoretical and numerical analysis of the twofactor rough Bergomi model and provide insights on the interplay between the different parameters for joint SPXVIX smile calibration.

Optimization Models for Autonomous Transfer Hub Networks ; Autonomous trucks are expected to fundamentally transform the freight transportation industry. In particular, Autonomous Transfer Hub Networks ATHN, which combine autonomous trucks on middle miles with humandriven on the first and last miles, are seen as the most likely deployment pathway of this technology. This paper presents three methods to optimize ATHN operations and compares them a constraintprogramming model, a columngeneration approach, and a bespoke network flow method. Results on a real case study indicate that the network flow model is highly scalable and outperforms the other two approaches by significant margins.

Decomposing LIBOR in Transition Evidence from the Futures Markets ; Applying historical data from the USD LIBOR transition period, we estimate a joint model for SOFR, Fed Funds, and Eurodollar futures rates as well as spot USD LIBOR and term repo rates. The framework endogenously models basis spreads between each of the benchmark rates and allows for the decomposition of spreads. Modelling the LIBOROIS spread as credit and fundingliquidity rollover risk, we find that the spike in the LIBOROIS spread during the onset of COVID19 was mainly due to credit risk, while on average credit and fundingliquidity risk contribute equally to the spread.

Clocks and trajectories in quantum cosmology ; We consider a simple cosmological model consisting of an empty Bianchi I Universe, whose Hamiltonian we deparametrise to provide a natural clock variable. The model thus effectively describes an isotropic universe with an induced clock given by the shear. Quantising this model, we obtain various different possible bouncing trajectories semiquantum expectation values on coherent states or obtained by the de BroglieBohm formulation and explicit their clock dependence, specifically emphasising the question of symmetry across the bounce.

Leptophilic New Physics and the Cabibbo Angle Anomaly ; The Cabibbo Angle Anomaly, an apparent deficit in firstrow CKM unitarity, can be addressed by leptophilic Standard Model extensions that generate new contributions to the Fermi constant and affect the determination of the CKM element Vud. We focus on simplified models with this property, including the Standard Model extended by vectorlike leptons, by the singly charged scalar singlet, or by a leptophilic Z' boson.

A radiative seesaw model in a supersymmetric modular A4 group ; We propose a supersymmetric radiative seesaw model with modular A4 symmetry. Thanks to contributions of supersymmetric partners to oneloop diagrams generating neutrino masses, we successfully fit neutrino data and obtain predictions in case of normal hierarchy in a minimal framework that would not be realized in a nonsupersymmetric model. We show a several predictive figures and demonstrate a best fit benchmark point through chi2

Taming modeling uncertainties with Mass Unspecific Supervised Tagging ; We address the modeling dependence of jet taggers built using the method of Mass Unspecific Supervised Tagging, by using two different parton showering and hadronisation schemes. We find that the modeling dependence of the results estimated by using different schemes in the design of the taggers and applying them to the same type of data is rather small, even if the jet substructure varies significantly between the two schemes. These results add great value to the use of generic supervised taggers for new physics searches.

SentimentAware Automatic Speech Recognition pretraining for enhanced Speech Emotion Recognition ; We propose a novel multitask pretraining method for Speech Emotion Recognition SER. We pretrain SER model simultaneously on Automatic Speech Recognition ASR and sentiment classification tasks to make the acoustic ASR model more emotion aware''. We generate targets for the sentiment classification using texttosentiment model trained on publicly available data. Finally, we finetune the acoustic ASR on emotion annotated speech data. We evaluated the proposed approach on the MSPPodcast dataset, where we achieved the best reported concordance correlation coefficient CCC of 0.41 for valence prediction.

Leptogenesis in Majoron Models without Domain Walls ; The emergence of domain walls is a wellknown problem in Majoron models for neutrino mass generation. Here, we present extensions of the Majoron model by righthanded doublets and triplets that prevent domain walls from arising. These extensions are highly interesting in the context of Leptogenesis as they impact the conversion of a lepton asymmetry to a baryon asymmetry by Sphaleron processes.

FiNCAT Financial Numeral Claim Analysis Tool ; While making investment decisions by reading financial documents, investors need to differentiate between inclaim and outofclaim numerals. In this paper, we present a tool which does it automatically. It extracts context embeddings of the numerals using one of the transformer based pretrained language model called BERT. After this, it uses a Logistic Regression based model to detect whether the numerals is inclaim or outofclaim. We use FinNum3 English dataset to train our model. After conducting rigorous experiments we achieve a Macro F1 score of 0.8223 on the validation set. We have opensourced this tool and it can be accessed from httpsgithub.comsohomghoshFiNCATFinancialNumeralClaimAnalysisTool

Distal systems in topological dynamics and ergodic theory ; We generalize a result of Lindenstrauss on the interplay between measurable and topological dynamics which shows that every separable ergodic measurably distal dynamical system has a minimal distal model. We show that such a model can, in fact, be chosen completely canonically. The construction is performed by going through the FurstenbergZimmer tower of a measurably distal system and showing that at each step, there is a simple and canonical distal minimal model. This hinges on a new characterization of isometric extensions in topological dynamics.

Geometric Digital Twinning of Industrial Facilities Retrieval of Industrial Shapes ; This paper devises, implements and benchmarks a novel shape retrieval method that can accurately match individual labelled point clusters instances of existing industrial facilities with their respective CAD models. It employs a combination of image and point cloud deep learning networks to classify and match instances to their geometrically similar CAD model. It extends our previous research on geometric digital twin generation from point cloud data, which currently is a tedious, manual process. Experiments with our joint network reveal that it can reliably retrieve CAD models at 85.2 accuracy. The proposed research is a fundamental framework to enable the geometric Digital Twin gDT pipeline and incorporate the real geometric configuration into the Digital Twin.

Modelling Underwater Acoustic Propagation using Oneway Wave Equations ; The primary contribution of this paper is to characterize the propagation of acoustic signal carrying information through any medium and the interaction of the travelling acoustic signal with the surrounding medium. We will use the concept of damped harmonic oscillator to model the medium and Milne's oscillator technique to map the interaction of the acoustic signal with the medium. The acoustic signal itself will be modelled using the oneway wave equation formulated in terms of acoustic pressure and velocity of acoustic waves through the medium. Using the abovementioned concepts, we calculated the effective signal strength, phase shift and time period of the communicated signal. Numerical results are generated to present the evolution of signal strength and received signal envelope in underwater environment.

Reciprocity in Machine Learning ; Machine learning is pervasive. It powers recommender systems such as Spotify, Instagram and YouTube, and healthcare systems via models that predict sleep patterns, or the risk of disease. Individuals contribute data to these models and benefit from them. Are these contributions outflows of influence and benefits inflows of influence reciprocal We propose measures of outflows, inflows and reciprocity building on previously proposed measures of training data influence. Our initial theoretical and empirical results indicate that under certain distributional assumptions, some classes of models are approximately reciprocal. We conclude with several open directions.

PETCI A Parallel English Translation Dataset of Chinese Idioms ; Idioms are an important language phenomenon in Chinese, but idiom translation is notoriously hard. Current machine translation models perform poorly on idiom translation, while idioms are sparse in many translation datasets. We present PETCI, a parallel English translation dataset of Chinese idioms, aiming to improve idiom translation by both human and machine. The dataset is built by leveraging human and machine effort. Baseline generation models show unsatisfactory abilities to improve translation, but structureaware classification models show good performance on distinguishing good translations. Furthermore, the size of PETCI can be easily increased without expertise. Overall, PETCI can be helpful to language learners and machine translation systems.

An ESBGK model for diatomic gases with correct relaxation rates for internal energies ; We propose a new ESBGK model for diatomic gases which allows for translationalrotational and translationalvibrational energy exchanges, as given by LandauTeller and Jeans relaxation equations. This model is consistent with the general definition of the vibrational and rotational collision numbers that are also commonly used in DSMC solvers. It is proved to satisfy the Htheorem and to give the correct transport coefficients, up to the volume viscosity.

Dependence structure for the product of bidimensional finitevariance VAR1 model components. An application to the cost of electricity load prediction errors ; In this paper we analyze the product of bidimensional VAR1 model components. For the introduced time series we derive general formulas for the autocovariance function and study its properties for different cases of crossdependence between the VAR1 model components. The theoretical results are then illustrated in the simulation study for two types of bivariate distributions of the residual series, namely the Gaussian and Student's t. We also show a possible practical application of the obtained results based on the data from the electricity market.

Beam Search for Feature Selection ; In this paper, we present and prove some consistency results about the performance of classification models using a subset of features. In addition, we propose to use beam search to perform feature selection, which can be viewed as a generalization of forward selection. We apply beam search to both simulated and realworld data, by evaluating and comparing the performance of different classification models using different sets of features. The results demonstrate that beam search could outperform forward selection, especially when the features are correlated so that they have more discriminative power when considered jointly than individually. Moreover, in some cases classification models could obtain comparable performance using only ten features selected by beam search instead of hundreds of original features.

Improved Analysis of CurrentSteering DACs Using Equivalent Timing Errors ; Currentsteering CS digitaltoanalog converters DACs generate analog signals by combining weighted current sources. Ideally, the current sources are combined at each switching instant simultaneously. However, this is not true in practice due to timing mismatch, resulting in nonlinear distortion. This work uses the equivalent timing error model, introduced by previous work, to analyze the signaltodistortion ratio SDR resulting from these timing errors. Using a behavioral simulation model we demonstrate that our analysis is significantly more accurate than the previous methods. We also use our simulation model to investigate the effect of timing mismatch in partiallysegmented CSDACs, i.e., those comprised of both equallyweighted and binaryweighted current sources.

Learning Stabilizable Deep Dynamics Models ; When neural networks are used to model dynamics, properties such as stability of the dynamics are generally not guaranteed. In contrast, there is a recent method for learning the dynamics of autonomous systems that guarantees global exponential stability using neural networks. In this paper, we propose a new method for learning the dynamics of inputaffine control systems. An important feature is that a stabilizing controller and control Lyapunov function of the learned model are obtained as well. Moreover, the proposed method can also be applied to solving HamiltonJacobi inequalities. The usefulness of the proposed method is examined through numerical examples.

Prediction of chaotic attractors in quasiperiodically forced logistic map using deep learning ; We forecast two different chaotic dynamics of the quasiperiodically forced logistic map using the wellknown deep learning framework Long ShortTerm Memory. We generate two data sets and use one in the training process and the other in the testing process. The predicted values are evaluated using the metric called Root Mean Square Error and visualized using the scatter plots. The robustness of the Long ShortTerm Memory model is evaluated using the number of units in the layers of the model. We also make multistep forecasting of the considered system. We show that the considered Long ShortTerm Memory model performs well in predicting chaotic attractors upto three steps.

Synthesis of memristive oneport circuits with piecewisesmooth characteristics ; A generalized approach for the implementation of memristive twoterminal circuits with piesewisesmooth characteristics is proposed on the example of a multifunctional circuit based on a transistor switch. Two versions of the circuit are taken into consideration an experimental model of the piecewisesmooth memristor Chua's memristor and a piecewisesmooth memristive capacitor. Physical experiments are combined with numerical modelling of the discussed circuit models. Thus, it is demonstrated that the considered circuit is a flexible solution for synthesis of a wide range of memristive systems with tuneable characteristics.

When Bousfield localizations and homotopy idempotent functors meet again ; We adopt semimodel categories to extend fundamental results related to Bousfield localizations of model categories. More specifically, we generalize BousfieldFriedlander Theorem and Hirschhorn Localization Theorem of cellular model categories to settings where their classical formulation does not apply. We use such results to answer, in the world of semimodel categories, an open problem posed by MayPonto about the existence of Bousfield localizations for Hurewicz and mixed type model structures.

Finite Sample Inference in Incomplete Models ; We propose confidence regions for the parameters of incomplete models with exact coverage of the true parameter in finite samples. Our confidence region inverts a test, which generalizes Monte Carlo tests to incomplete models. The test statistic is a discrete analogue of a new optimal transport characterization of the sharp identified region. Both test statistic and critical values rely on simulation drawn from the distribution of latent variables and are computed using solutions to discrete optimal transport, hence linear programming problems. We also propose a fast preliminary search in the parameter space with an alternative, more conservative yet consistent test, based on a parameter free critical value.