Transfer Learning for Human & AI

Colonoscopy is a procedure to detect colorectal polyps which are the primary cause for developing colorctal cancer . Polyp segmentation is a challenging task due to the diverse shape, size, color, and texture of polyps . Instead of directly adding encoder features to the respective decoder layer, we introduce anAdaptive Global Context Module (AGCM) which focuses only on the encoder’ssignificant and hard fine-grained features .…

The weights of a deep neural network model are optimized in conjunction with the governing flow equations to provide a model for sub-grid-scale stresses in a turbulent jet at Reynolds number $Re_0=6\,000$. In-sample and out-of-sample testing on multipledual-jet configurations show that its required mesh density in each coordinatedirection for prediction of mean flow, Reynolds stresses, and spectra is half that needed by the dynamic Smagorinsky model for comparable accuracy .…

Recent advances in reinforcement learning have inspired increasing interest in learning user modeling adaptively through dynamic interactions . We propose a generative inversereinforcement learning for user behavioral preference modelling . Instead of using predefined reward functions, our model canautomatically learn the rewards from user’s actions based on discriminativeactor-critic network and Wasserstein GAN .…

This paper presents a robust controller using a Linear Parameter Varying(LPV) model of the lane-keeping system with parameter reduction . Multiple varying parameters lead to a high number of scheduling variables and cause massive computational complexity . We designed the LPV robust feedback controller using the reducedmodel solving a set of Linear Matrix Inequality (LMI) The effectiveness of the proposed system is validated with full vehicle dynamics from CarSim on aninterchange road .…

In this letter, we investigate the stability properties of a discrete-timemultilinear system . We establish theoretical results on the criteria fordetermining the internal stability of multilinare systems via tensor spectraltheory . In particular, we show that the Z-eigenvalues of the dynamic tensorplay a significant role in the stability analysis .…

Location data are widely used in mobile apps, ranging from location-basedrecommendations, to social media and navigation . But serious privacy concerns arise if users share their locationhistory with the service provider in plaintext . The underlying searchable encryption primitives required to perform the matching on ciphertexts are expensive, and without a proper encoding oflocations and search predicates, the performance can degrade a lot .…

The simplicity of the visual servoing approach makes it an attractive option for tasks dealing with vision-based control of robots in many real-world applications . Attaining precise alignment for unseen environments pose a challenge to existing visual servoational approaches . The recent data-driven approaches face issues when generalizing to novel environments .…

We consider a variant of the channel simulation problem with a single input and multiple outputs . We show that the growth rate of theexpected codeword length is sub-linear in $n$ when a power law bound is satisfied . An application of multiple-outputs channel simulation is thecompression of probability distributions .…

Factorized layers–operations parameterized by products of two or morematrices–occur in a variety of deep learning contexts . We study how to initialize and regularize deepnets containing such layers . We highlight the benefits of spectralinitialization and Frobenius decay . In modelcompression, we show that they enable low-rank methods to significantlyoutperform both unstructured sparsity and tensor methods on the task of training low-memory residual networks .…

Despite decades of engineering and scientific research efforts, separation of concerns in software development remains not fully achieved . We show that business-domain codeinscriptions play an even larger role in this challenge than the crosscutting of concerns phenomenon . We introduce a new methodology, called MetadataInterpretation Driven Development (MIDD) that suggests a possible path to .…

We consider the task of approximating the ground state energy of two-localquantum Hamiltonians on bounded-degree graphs . Most existing algorithmsoptimize the energy over the set of product states . Here we describe a family of shallow quantum circuits that can be used to improve the approximation ratioachieved by a given product state .…

Signal processing and communication communities have witnessed the rise of exciting communication technologies in recent years . We discuss the recent developments in the field and put forward promising candidates for future research and development . We also envision an ultimate RIS architecture, which is able to adjust its operation modes dynamically, and introduce the new concept of PHY slicing over RISs towards 6G wireless networks.…

A linear dynamical system is said to beasymptotically stable if all of the eigenvalues of A lie in the open left-half of the complex plane . In practice, stability of a system can be violated by arbitrarily small perturbations such as modeling errors .…

Music is becoming an essential part of daily life . There is an urgent need todevelop recommendation systems to assist people targeting better songs withfewer efforts . As the interactions between users and songs naturally construct a complex network, community detection approaches can be applied to reveal potential interests on songs .…

We construct a structure-preserving finite element method and time-steppingscheme for compressible barotropic magnetohydrodynamics . The method preserves thebalance laws governing the evolution of total energy and magnetic helicity, and preserves mass and the constraint $ \operatorname{div}B = 0$ to machine precision .…

In legged locomotion, the relationship between different gait behaviors and energy consumption must consider the full-body dynamics and robot control as a whole . We developed a 2-phasenonlinear optimization pipeline for dynamic stepping, which generatesreachability maps showing complex energy-stepping relations .…

Size and distance perception in Virtual Reality have been widelystudied in a controlled laboratory setting with a small number of participants . We describe a fully remote perceptual study with a gamified protocol to encourage participant engagement . Varying eye-height from the floor plane showed no significant effect on the judgements .…

Most methods for publishing data with privacy guarantees introduce randomness, which reduces the utility of the published data . In this paper, we study the privacy-utility tradeoff by taking maximal leakage as the privacy measure and the expected Hamming distortion as the utility measure .…

In a recent paper, a quasi-nonlocal coupling method was introduced to seamlessly bridge a nonlocal diffusion model with the classical localdiffusion counterpart in a one-dimensional space . The proposed couplingframework removes inconsistency on the interface, preserves the balance offluxes, and satisfies the maximum principle of diffusion problem .…

The digital transformation is turning archives, both old and new, into data . Automation in the form of artificial intelligence techniques is increasingly applied both to scale traditional recordkeeping activities, and to experiment with novel ways to capture, organise and access records .…

This paper proposes new practical design tools for the robust motion controlsystems based on disturbance observer (DOB) Although DOB has long been used in several motion control applications, it has insufficient analysis and designtools . The paper proposes a new practical robustness constraint, which improve the robustness at high frequencies, on the bandwidth of a DOB and nominalinertia .…

Beldiceanu et al.presented a generic filtering scheme for global constraints based on graph properties . This scheme strongly relies on the computation of graph properties’bounds and can be used in the context of graph domain variables and constraints . Bounds of MIN_NCC had been defined for the graph-based representation of global constraint for the path_with_loops graphclass .…

We present an in-house modelling framework for Ferroelectric TunnellingJunctions (FTJ) and an insightful study of the design of FTJs as synapticdevices . Results show that a moderately low-k tunnelling dielectric (e.g. SiO2) can increase the read current and the current dynamic range .…

Cultural differences influence user preferences and interaction methods . We believe that it is equally important to apply this inquiry to digital accessibility and how accessibility fits within the design process around the world . We hope that this inquiry will also be applied to how digital accessibility fits into design process .…

An efficient real-time convex optimized guidance and control strategy is presented . Successive convexification of the parafoil guidanceproblem guarantees local optimality . Exhaustive Monte-Carlo simulations show performanceimprovements of about one order of magnitude . The concept proposed is simple,yet general, as it scales to any atmospheric parfoil landing system and allow efficient implementation relying only on the turn rate information .…

We present a new method to construct Hermitian self-orthogonal $[n,k,d]_{q^2} codes with large dimensions $k\frac{n+q-1}{q+1$ We provide entanglement-assisted quantum error-correcting codes with new parameters . We prove that any Hermitan self-or-hermitian code gives riseto an $ [n]k, d]_q# code with a $[ n,k]__k-1$ code with $0\le\ell \le k-1$.…

To implement fair machine learning in a sustainable way, choosing the rightfairness objective is key . The most appropriate fairness definition for an artificial intelligence system is a matter of ethical standards and legal requirements . In thisposition paper, we propose to use a decision tree as means to explain andjustify the implemented kind of fairness to the end users .…

Reachability analysis of nonlinear dynamical systems is a challenging andcomputationally expensive task . Computing the reachable states for linear systems, in contrast, can often be done efficiently in high dimensions . The Koopman operator links the behaviors of a nonlinear system to a linear system embedded in a higher dimensional space, with an additional set of so-calledobservable variables .…

We present a framework that enables the discovery of diverse andnatural-looking motion strategies for athletic skills such as the high jump . The strategies are realized as control policies for physics-based characters . We propose a Pose Variational Autoencoder(P-VAE) to constrain the actions to a subspace of natural poses .…

CMA-Net consists of two novel mutual attention modules aiming at fusing the high level features from the modalities of all-in-focus and depth . It can run at a speed of 53 fps, thus being much faster than the state-of-the-art multi-modal SOD methods .…

We investigate artificial compressibility (AC) techniques for the timediscretization of the incompressible Navier-Stokes equations . AC timestepping uncouples at each time step thevelocity update from the pressure update . We consider both first-order and second-order time schemes and either an implicit or an explicit treatment of the nonlinear convection term .…

We consider measuring the protocol leakage by the log-ratio distance (which was popularized by its use in the differential privacy framework) We show that a protocol with (noticeable) accuracy can betransformed into an OT protocol . Our results hold for both theinformation theoretic and the computational settings, and can be viewed as a”fine grained” approach to “weak OT amplification” We then use the above result to fully characterize the complexity ofdifferentially private two-party computation for the XOR function .…

We extend the Transformer network to attend the most relevant information withrespect to the target speaker given the characteristics of his or her voices as a form of contextual information . The idea has a natural interpretation interms of the selective attention theory .…

This paper aims to better understand the link between aperiodicity in subshifts and pattern complexity . We prove aquadratic lower bound on their pattern complexity for a class of substitutive sub-shifts . We also prove that the recent bound ofKari and Moutot, showing that any aperiodic subshift has pattern complexity atleast $mn+1$, is optimal for fixed $m$ and $n$.…

In this paper, we investigate System Fat, or atomic System F, a weak predicative fragment of System F whose typable terms coincide with the simply typable ones . We show that the type-checking problem for Fat is decidable and we propose an algorithm which sheds some new light on the source of undecidability in full System F .…

In computational PDE-based inverse problems, a finite amount of data is collected to infer unknown parameters in the PDE . The best, and often infeasible, scenario is when the full input-to-output (ItO) map is available: This is the typical setting in many theoretical inverse problems .…

The proposed method is generic enough to be integrated to many variants of weighted ICP without adaptation . It supports localization robustness in harsher terrains, says the authors . The method decreasesthe error by 9.26% in translation and by 21.84% in rotation.…

Stakeholders make various types of decisions with respect to requirements, design, management, and so on during the software development life cycle . These decisions are typically not well documented and classified due to limited human resources, time, and budget . In this paper, we aimed at automaticallyclassifying decisions into five types to help stakeholders better document and understand decisions .…

Unsupervised Document Expansion with Generation (UDEG) generates diversesupplementary sentences for the original document without using labels on query-document pairs for training . We validate our framework on two standard IR benchmarkdatasets . The results show that our framework significantly outperformsrelevant expansion baselines for IR.…

This paper envisions a multi-agent system for detecting the presence of hatespeech in online social media platforms such as Twitter and Facebook . We introduce a novel framework employing deep learning techniques to coordinatethe channels of textual and im-age processing .…

Building human-like agents aims to learn and think like human intelligence . We propose a novelVR-based toolkit, VECA, which enables building fruitful virtual environments . VECA provides a humanoid agentand an environment manager, enabling the agent to receive rich human-Likeperception and perform comprehensive interactions .…

We present a novel class of high-order space-time finite element schemes for the Poisson-Nernst-Planck (PNP) equations . We prove that our schemes are massconservative, positivity preserving, and unconditionally energy stable for any order of approximation . This is the first class of (arbitrarily) high order accurate schemes for PNP equations that simultaneously achieve all these three properties .…

Recurrent Neural networks (RNN) have shown promising potential for learningdynamics of sequential data . Artificial neural networks are known to exhibit poor robustness in presence of input noise . The variance of the output of the noisy system is adopted as a robustness measure to quantify the impact of noise on learning .…

Volumetric datasets have captured high-quality 3Dvideo, but accompanying audio is often neglected . The presented volumetric dataset was constructed from audio and videostudio recordings of scenes to yield forty short action sequences . Potentialuses in technical and scientific tests are discussed.…

LACLICHEV is based on data collections’ exploration techniques combined with informationretrieval, data analytics, and geographic querying and visualization . Thisenvironment provides tools for curating, exploring and analyzing historical articles, their description and location, and the vocabularies used to refer to meteorological events .…

We introduce generalized spatially coupled parallel concatenated codes (GSC-PCCs) as a class of turbo-like turbo-style codes . We show that the proposed codes have some niceproperties such as threshold saturation and that their decoding thresholdsimprove with the repetition factor $q$. We also suggest that the codes asymptotically approach the capacity as $q$ tends toinfinity with any given constituent convolutional code .…

We consider decomposing the information content of three Gaussian random vectors using the partial information decomposition(PID) framework . Barrett previously characterized the Gaussian PID for a scalar”source” or “message” in closed form – we extend this to the case where themessage is a vector .…

Enterprise modeling deals with the increasing complexity of processes and systems by operationalizing model content and by linking complementary modelsand languages . To enable this amplification and turn models intocomputer-processable structures a comprehensive formalization is needed . Thispaper presents a generic formalism based on typed first-order logic and provides a perspective on the potential and benefits arising for a variety of research issues in conceptual modeling .…

We consider the uplink transmission between a multi-antenna ground station and an unmanned aerial vehicle (UAV) swarm . The UAVs are assumed as intelligent agents, which can explore their optimal three dimensional (3-D) deployment tomaximize the channel capacity of the multiple input multiple output (MIMO)system .…

We show that it is decidable whether two regular languages of infinite trees are separable by a deterministic language . We considertwo variants of separability, depending on whether the set of priorities of theseparator is fixed, or not . In each case, we show that separability can bedecided in EXPTIME, and that separating automata of exponential size suffice .…