## Return migration of German affiliated researchers Analyzing departure and return by gender cohort and discipline using Scopus bibliometric data 1996 2020

The international migration of researchers is a highly prized dimension of scientific mobility and motivates considerable policy debate . In this study, we use Scopus bibliometric data on 8 million publications from 1.1 million researchers who have published at least once with an affiliation address from Germany in 1996-2020 .…

## Communicating Patient Health Data A Wicked Problem

Designing patient-collected health data visualizations to support discussing patient data during clinical visits is a challenging problem due to theheterogeneity of the parties involved . Designers must ensure that all parties’ needs are met . This complexity makes it challenging to find a definitive solution that can work forevery individual .…

## BPPChecker An SMT based Model Checker on Basic Parallel Processes

Basic Parallel Process (BPP), as a subclass of Petri nets, can be used as a model for describing and verifying concurrent programs with lowercomplexity . We propose and implement BPPChecker, the first model checker forverifying CTL on BPP . For EF operator, we reduce the model checking of EF-formulas to the satisfiabilityproblem of existential Presburger formula .…

## Adaptive and Fair Transformation for Recoverable Mutual Exclusion

Mutual exclusion is one of the most commonly used techniques to handlecontention in concurrent systems . Traditionally, mutual exclusion algorithms have been designed under the assumption that a process does not fail while acquiring/releasing a lock or executing its critical section .…

## Cascaded Fast and Slow Models for Efficient Semantic Code Search

The goal of natural language semantic code search is to retrieve asemantically relevant code snippet from a fixed set of candidates using anatural language query . Existing approaches are neither effective nor efficient enough towards a practical semanticcode search system .…

## Accelerating Genetic Programming using GPUs

Genetic Programming (GP), an evolutionary learning technique, has multiple applications in machine learning such as curve fitting, data modelling, featureselection, classification etc. GP has several inherent parallel steps, making it an ideal candidate for GPU based parallelization . We evaluate our algorithm on synthetic datasets for the PagiePolynomial (ranging in size from $4096$ to $16$ million points) We run performance benchmarks on our algorithmand gplearn, profiling the training time, test accuracy, and loss .…

## A Broad Spectrum Diffractive Network via Ensemble Learning

A broad-spectrum diffractive deep neural network (BS-D2NN) incorporates multi-wavelength channels of input lightfields and performs a parallel phase-only modulation . A complementary multi-channel base learner cluster is formed in ahomogeneous ensemble framework based on the diffractive dispersion duringlightwave modulation . The BS-D1NN can be trained usingdeep learning algorithms so as to perform a kind of wavelength-insensitivehigh-accuracy object classification .…

## Zipping Strategies and Attribute Grammars

Zippers provide a simple, but generic tree-walk mechanism that is the building blocktechnique we use to express the purely-functional embedding of both techniques . The combined embedding is easier to maintain and extend since it is written in a concise and uniformsetting .…

## Optimal Distribution Design for Irregular Repetition Slotted ALOHA with Multi Packet Reception

ALOHA (IRSA) holds a great potential in improving the accesscapacity of massive machine type communication systems . K = 2 (multi-packet reception capability)may be the most suitable scheme for scenarios that allow smaller resourceefficiency in exchange for greater throughput . In this paper, we analyticallyderive an optimal transmission probability distribution for IRSA with K =2 .…

## Low rank Matrix Recovery With Unknown Correspondence

We study a matrix recovery problem with unknown correspondence . We show that it is possible to recover $M$ via solving a nuclear normminimization problem . We propose an algorithm,$\text{M}^3\text {O}$ (Matrix recovery via Min-Max Optimization) which recastststhis combinatorial problem as a continuous minimax optimization problem andsolves it by proximal gradient with a Max-Oracle .…

## Development of Quantum Circuits for Perceptron Neural Network Training Based on the Principles of Grover s Algorithm

The demonstrated quantum circuits were based on the principles of Grover’s Search Algorithm . Theperceptron was chosen as the architecture for the example neural network . Themultilayer perceptron is a popular neural network architecture due to its applicability for solving a wide range of problems .…

## Return migration of German affiliated researchers Analyzing departure and return by gender cohort and discipline using Scopus bibliometric data 1996 2020

The international migration of researchers is a highly prized dimension of scientific mobility and motivates considerable policy debate . In this study, we use Scopus bibliometric data on 8 million publications from 1.1 million researchers who have published at least once with an affiliation address from Germany in 1996-2020 .…

## The Pebble Relation Comonad in Finite Model Theory

The pebbling comonad, introduced by Abramsky, Dawar and Wang, provides acategorical interpretation for k-pebble games from finite model theory . The coKleisli category of the pebble-relation comonads specifies equivalences under different fragments and extensions of infinitary k-variable logic . We prove a new Lov\’asz-type theorem relating pathwidth to therestricted conjunction fragment of k- variable logic with counting quantifiers .…

## Improving overlay maps of science combining overview and detail

Overlay maps of science are global base maps over which subsets of publications can be projected . Such maps can be used to monitor, explore, and study research through its publication output . The aim of this study is to improve overlaymaps of science to provide both features in a single visualization .…

## Intent based Product Collections for E commerce using Pretrained Language Models

Building a shopping product collection has been primarily a human job . With the manual efforts of craftsmanship, experts collect related but diverse products with common shopping intent that are effective when displayed together . Automatically constructing a collection requires an ML system to learn a complex relationship between the customer’s intent and the product’sattributes .…

## A Static Analysis Framework for Data Science Notebooks

Notebooks provide an interactive environment for programmers to develop code, analyze data and inject interleaved visualizations in a single environment . A major pitfall that data scientists encounter is unexpected behaviour caused by the unique out-of-order execution model of notebooks .…

## Faster Modular Composition

A new Las Vegas algorithm is presented for the composition of two polynomialsmodulo a third one, over an arbitrary field . When the degrees of thesepolynomials are bounded by $n, the algorithm uses$O(n^{1.43)$field operations . The previous fastest algebraic algorithms, due to Brent and Kung in 1978, require$ O(n.63) field operations in general, and ${n.3/2+o(1)$ in particular case of power series over a field of large enoughcharacteristic .…

## Least Squares on GPUs in Multiple Double Precision

This paper describes the application of the code generated by the CAMPARYsoftware to accelerate the solving of linear systems in the least squares sense . The goal is to use accelerators to offset the cost overheadcaused by multiple double precision arithmetic .…

## Dynamics of Cross Platform Attention to Retracted Papers Pervasiveness Audience Skepticism and Timing of Retractions

Retracted papers often circulate widely on social media, online news outlets and other websites before their official retraction . The spread of potentiallyinaccurate or misleading results from retracted papers can harm the scientific community and the public . This finding indicates that untrustworthyresearch penetrates even curated platforms and is often shared uncritically,plplifying the negative impact on the public, say the authors .…

## Uplink Power Control in Integrated Access and Backhaul Networks

Integrated access and backhaul (IAB) network is a novel radio access network(RAN) solution, enabling network densification for 5G and beyond . We use power control combined with resource allocation algorithms todevelop efficient IAB networks with high service coverage . We also study the effect of different parameters including minimum data raterequirement, coverage distance and transmit power on the network performance .…

## Hindsight Posterior guided training of retrievers for improved open ended generation

Many text generation systems benefit from using a retriever to retrieve passages from a textual knowledge corpus . For open-ended generation tasks, many varied passages may be equally relevant . We propose using an additional guide retriever that is allowed to use the target output and “inhindsight” retrieve relevant passages during training .…

## New techniques for bounding stabilizer rank

In this work, we present number-theoretic and algebraic-geometric techniques for bounding the stabilizer rank of quantum states . First, we find the first non-trivial examples of quantumstates with multiplicative stabilizer ranks under the tensor product . We also find new bounds on the genericstabilizer rank.…

## Peer reviewers equally critique theory method and writing with limited effect on manuscripts content

Peer review aims to detect flaws and deficiencies in the design andinterpretation of studies . However, it has been questioned whether peer review fulfils this function . Studies have highlighted a stronger focus of reviewers on critiquingmethodological aspects of studies and the quality of writing in biomedicalsciences .…

## Using Psychological Characteristics of Situations for Social Situation Comprehension in Support Agents

Support agents that help users in their daily lives need to take into account not only the user’s characteristics, but also the social situation of the user . Research shows that it is important to alsodetermine the meaning of a situation, a step which we refer to as socialsituation comprehension .…

## A framework for the analysis of fully coupled normal and tangential contact problems with complex interfaces

An extension to the interface finite element with eMbedded Profile for JointRoughness is herein proposed for solving the contact problem between a rigid indenter of any complex shape and anelastic body under generic oblique load histories . The actual shape of theindenter is accounted for as a correction of the gap function .…

## Tight Lipschitz Hardness for Optimizing Mean Field Spin Glasses

We study the problem of algorithmically optimizing the Hamiltonian $H_N/N of aspherical or Ising mixed$p$-spin glass . We prove that for mixed even$p-spin models, no algorithm satisfying anoverlap concentration property can produce an objective larger than $\mathsf{ALG$ with non-negligible probability .…

## StreaMulT Streaming Multimodal Transformer for Heterogeneous and Arbitrary Long Sequential Data

This paper tackles the problem of processing and combining efficientlyarbitrary long data streams . Common applications can be, for instance, long-timeindustrial or real-life systems monitoring from multimodal heterogeneous data . To tackle this problem, we propose StreaMulT, a Streaming Multimodal Transformer, relying on cross-modalattention and an augmented memory bank to process arbitrary long inputsequences at training time and run in a streaming way at inference .…

## Weakly Supervised Semantic Segmentation by Pixel to Prototype Contrast

Authors propose novel pixel-to-prototype contrast regularization terms that are conducted cross different views and within per single view of animage . They adopt two sample mining strategies, namedsemi-hard prototype mining and hard pixel sampling, to better leverage hardexamples while reducing incorrect contrasts caused due to the absence of pixel-wise labels .…

## Brittle interpretations The Vulnerability of TCAV and Other Concept based Explainability Tools to Adversarial Attack

Methods for model explainability have become increasingly critical fortesting the fairness and soundness of deep learning . In safety-critical applications, there is need for security around not only the machine learning pipeline but also the modelinterpretation process . We show that by perturbing the examples of the concept that is being investigated, we can radically change the output of the interpretability method, e.g.…

## Offline Reinforcement Learning with Soft Behavior Regularization

Most prior approaches to offline reinforcement learning (RL) utilize behavior regularization, typically augmenting existing off-policyactor critic algorithms with a penalty measuring divergence between the policy and the offline data . We propose a practical way to compute the density ratio and demonstrate its equivalence to a state-dependentbehavior regularization .…

## Coarse to Fine Video Retrieval before Moment Localization

Current state-of-the-art methods for video corpus moment retrieval (VCMR) often use similarity-based feature alignment approach . However, late fusion methods like cosine similarityalignment are unable to make full use of the information from both query texts and videos . In this paper, we combine feature alignment with feature fusion topromote the performance on VCMR .…

## Algebraic Reasoning of Quantum Programs via Non Idempotent Kleene Algebra

We investigate the algebraic reasoning of quantum programs inspired by the success of classical program analysis based on Kleene algebra . We propose the Non-idempotent Kleena Algebra (NKA) as a natural alternative and identify complete and sound semantic models for NKA as well astheir appropriate quantum interpretations .…

## The Neural MMO Platform for Massively Multiagent Research

Neural MMO is a computationally accessible research platform that combines large agent populations, long time horizons, open-ended tasks, and modular gamesystems . We present Neural MMO as free and opensource software with active support, ongoing development, documentation, and training, logging, and visualization tools to help users adapt to the new setting .…

## Parallel Algebraic Effect Handlers

Algebraic effects and handlers support composable and structured control-flowabstraction . We formalize an untyped lambdacalculus which models two key features, effect handlers and parallelizablecomputations . We present various interesting examples expressible in our calculus, and provide a Haskell implementation . We hope this paper provides a basis for future designs and implementations of parallel algebraic effect handlers of this type of effect handlers .…

## Robust monolithic solvers for the Stokes Darcy problem with the Darcy equation in primal form

We construct mesh-independent and parameter-robust monolithic solvers for the primal Stokes-Darcy problem . Numerical experiments demonstrate the parameters of the proposed solvers . The suggested preconditioners utilize operators in fractionalSobolev spaces . In each case, robustpreconditioners are derived using a unified theoretical framework .…

## Adversarial examples by perturbing high level features in intermediate decoder layers

Instead ofperturbing pixels, we use an encoder-decoder representation of the input image and perturb intermediate layers in the decoder . This changes the high-levelfeatures provided by the generative model . We formulatethis task as an optimization problem by minimizing the Wasserstein distancebetween the adversarial and initial images under a misclassificationconstraint .…

## A CLIP Enhanced Method for Video Language Understanding

We propose aCLIP-Enhanced method to incorporate the image-text pretrained knowledge intodownstream video-text tasks . Combined with several other improved designs, ourmethod outperforms the state-of-the-art by $2.4\%$ ($57.58$ to $60.00$)…

## SGoLAM Simultaneous Goal Localization and Mapping for Multi Object Goal Navigation

SGoLAM is ranked 2nd in the CVPR 2021 MultiON(Multi-Object Goal Navigation) challenge . It does not require any training of neural networks, it could be used in an off-the-shelfmanner, and amenable for fast generalization in new, unseen environments . The mapping module converts observations into an occupancy map, and the goal localization modulemarks the locations of goal objects .…

## Provably Efficient Multi Agent Reinforcement Learning with Fully Decentralized Communication

Distributed exploration reduces samplingcomplexity in multi-agent RL (MARL) We investigate the benefits to performance in MARL when exploration is fully decentralized . We show that incorporating more agents and more information sharing into the group learning scheme speeds up convergence to the optimal policy .…

## Video based cattle identification and action recognition

Deep learning models have been developed and tested with videos acquired in a farm . An accuracy of 84.4\% has beenachieved for the detection of drinking events, and 94.4% for grazing events . We demonstrate a working prototype for the monitoring of cow welfare by analysing the animal behaviours of individual animals to enableautomated farm provenance .…

## Semantically Distributed Robust Optimization for Vision and Language Inference

Analysis of vision-and-language models has revealed their brittleness underlinguistic phenomena such as paraphrasing, negation, textual entailment, and textual substitutions with synonyms or antonyms . In this paper, we present a model-agnosticmethod that utilizes a set linguistic transformations in a distributed robustoptimization setting .…

## DeepSSM A Blueprint for Image to Shape Deep Learning Models

Statistical shape modeling (SSM) characterizes anatomical variations in apopulation of shapes generated from medical images . SSM requires consistentshape representation across samples in shape cohort . Theseshape representations are then used to extract low-dimensional shapedescriptors that facilitate subsequent analyses in different applications .…

## Region Semantically Aligned Network for Zero Shot Learning

Zero-shot learning (ZSL) aims to recognize unseen classes based on the knowledge of seen classes . Previous methods focused on learning directembeddings from global features to the semantic space in hope of knowledgetransfer from seen classes to unseen classes . Instead of using globalfeatures which are obtained by an average pooling layer after an image encoder, we directly utilize the output of the image .…

## Semi supervised Multi task Learning for Semantics and Depth

Multi-Task Learning (MTL) aims to enhance the model generalization by sharingrepresentations between related tasks for better performance . Typical MTLmethods are jointly trained with the complete multitude of ground-truths forall tasks simultaneously . However, one single dataset may not contain theannotations for each task of interest .…

## Unrolled Variational Bayesian Algorithm for Image Blind Deconvolution

In this paper, we introduce a variational Bayesian algorithm (VBA) for imageblind deconvolution . Our generic framework incorporates smoothness priors on the unknown blur/image and possible affine constraints (e.g., sum to one) on the blur kernel . One of our main contributions is the integration of VBA withina neural network paradigm, following an unrolling methodology .…

## Unsupervised Data Driven Nuclei Segmentation For Histology Images

An unsupervised data-driven nuclei segmentation method for histology images,called CBM, is proposed in this work . CBM consists of three modules applied in a block-wise manner: 1, 2, 3, incorporation ofgeometric priors with morphological processing . Experiments on the MoNuSeg dataset validate the effectiveness of the proposed CBM method .…

## Compressibility of Distributed Document Representations

We propose CoRe, a straightforward, representation learner-agnostic framework suitable for representation compression . The CoRe’s performance was studied on a collection of 17 real-life corpora from biomedical,news, social media, and literary domains . We explored the behavior whenconsidering contextual and non-contextual document representations, differentcompression levels, and 9 different compression algorithms .…

## A more direct and better variant of New Q Newton s method Backtracking for m equations in m variables

In this paper we propose a variant of New Q-Newton’s method Backtracking . The update rule of our method is $x\mapsto x-\gamma (x)w(x)$. Good theoretical guarantees are proven, in particular for systems ofpolynomial equations . In “generic situations”, we will also discuss a way to avoid that the limit of the constructed sequence is a solution of $H(x), but not of$F(x)=0\$ The limit is a .…

## A Comprehensive Study on Torchvision Pre trained Models for Fine grained Inter species Classification

Transfer Learning is an effective method of achieving extremely good performance with insufficient training data . This study aims to explore different pre-trained models offered in the PyTorch library . Torchvision package offers us many models to apply the Transfer Learning onsmaller datasets .…

## Nuisance Label Supervision Robustness Improvement by Free Labels

Nuisance-label Supervision (NLS) module makes models more robust to nuisance factor variations . Nuisance factors are those irrelevant to a task, and an ideal model should be invariant to them . Experiments show consistent improvement in robustness towards image corruption and appearance change in actionrecognition .…