Transfer Learning for Human & AI

SMLtoCoq provides an automatictranslation of SML programs and function contracts into Coq . Programs are translated into theorems, which can then be proved using the Equations plugin and other wellestablished Coq libraries . We provide a Coq version of many parts of the SML’s basislibrary, so that calls to these libraries are kept almost as is.…

We define a game on distributed Petri nets, where several players interact with each other, and with an environment . The players, or users, have perfect knowledge of the current state, and pursue a common goal . Such goal is expressed by Alternating-time Temporal Logic (ATL) The users have a winning strategy if they can cooperate to reach their goal, no matter how the environment behaves .…

This paper provides a compact method to lift the free exponentialconstruction of Mellies-Tabareauau-Tasson over the Hyland-Schalk double glueingfor orthogonality categories . A condition “reciprocity of orthogontality” is presented simply enough to lift free exponential over the double glueing interms of the orthogonsality .…

In applications such as natural language processing or computer vision, one is given a large $n \times d$ matrix $A = (a_{i,j)$ and would like to compute a matrix decomposition . A natural way to do this would be to simply apply $f$ to eachentry of $A$ and then compute the decomposition, e.g.,…

Homophily — the tendency of nodes to connect to others of the same type — is a central issue in the study of networks . Here we take a local view of homophily, defining notions of first-order homophilies of a node (its individualtendency to link to similar others) The gap in averages defies simple intuitiveexplanations, applies to globally heterophilous networks and isreminiscent of but structurally distinct from the Friendship Paradox .…

Energy Saver is a software that monitors electric energy consumption from data capture to consumption forecast for the following month . It uses Open Source technologies applied to the internet of Things (IoT), embedded systems, and Long Short-Term Memory NeuralNetworks (LSTM) The software has as its objective the monitoring of electricenergy consumption .…

Given a data table, dataworkers usually need to experience a tedious and time-consuming process to select meaningful combinations of data columns for creating charts . We contribute a deep-learning-based method that assists in designinganalytical dashboards . Our model could makerecommendations given optional user-input selections of data .…

We give the first single-pass streaming algorithm for Column Subset Selection . We study the $\ell_p$ norm loss since it is often considered more robust to noise than the standard Frobenius norm . The space complexity of the streaming algorithm is optimal up to a logarithmic factor .…

We present the cognitive architecture of an autonomous agent for active portfolio management in decentralized finance . It involves activities such as asset selection, portfolio balancing, liquidity provision, and trading . Partial implementation of the architecture is provided and supplied with preliminary results .…

How does information consumption affect behaviour in the context of theCOVID-19 pandemic? A popular hypothesis states that the so-called infodemics have substantial impact on orienting individual decisions . A competing hypothesis stresses that exposure to vast amounts of even contradictoryinformation has little effect on personal choices .…

Software Product Lines (SPLs) are families of related software products developed from a common set of artifacts . Most existing analysis tools can be applied to a single product at a time, but not to an entire SPL . In this paper, we take an existing declarative analysis (behaviouralteration) written in Grok, port it to Datalog, and apply it to a set of automotive software product lines from General Motors .…

We study a generalization of $k$-center clustering, first introduced byKavand et. al., where instead of one set of centers, we have two types of centers . Each red center is at least $alpha$distant from each blue center, and the goal is to minimize the covering radius .…

This paper presents a unified computational framework for estimation of distances, geodesics and barycenters of merge trees . We extend recent work on the edit distance [106] and introduce a new metric, called the Wassersteindistance between merge trees, which is purposely designed to enable efficientcomputations .…

Decompilers are useful tools for software analysis and support in the absence of source code . None of the existing tools support modern AMD GPU architectures such as AMD GCN and RDNA . We aim at developing the first assembly decompiler tool for a modernAMD GPU architecture that generates code in the OpenCL language .…

The Pareto-optimality criterion is based on the total support volume and the “secluded” support volume . Sincetotal support volume mainly depends on the build orientation and the secluded volume is dictated by the machining setup . The accessibilityanalysis relies on the inaccessibility measure field (IMF), which is acontinuous field in the Euclidean space .…

Homophily — the tendency of nodes to connect to others of the same type — is a central issue in the study of networks . Here we take a local view of homophily, defining notions of first-order homophilies of a node (its individualtendency to link to similar others) The gap in averages defies simple intuitiveexplanations, applies to globally heterophilous networks and isreminiscent of but structurally distinct from the Friendship Paradox .…

MultiBench is a systematic and unified large-scale benchmark spanning 15 datasets, 10 modalities, 20 prediction tasks, and 6 research areas . MultiBench provides an automated end-to-end machine learning pipeline that simplifies and standardizes data loading, experimental setup, and model evaluation .…

This paper proposes a library and scripts for automated generation of routine functions inthe Julia programming language for a set of numerical schemes of Runge-Kuttamethods . The proposed approach allows us to use asingle template for editing, instead of modifying each individual function to be compared .…

In the numerical linear algebra community, it is thought that to obtain nearly-optimal bounds for various problems such as rank computation and finding a linearly independent subset of columns, regression, low rankapproximation, maximum matching on general graphs and linear matroid union, onewould need to resolve the logarithmic factors in the sketching dimension for existingconstant factor approximation oblivious subspace embeddings .…

Type theories can be formalized using the intrinsically (hard) or theextrinsically (soft) typed style . In large libraries of type theoreticalfeatures, often both styles are present, which can lead to code duplication and integration issues . We define an operator that systematically translates a hard-typed into thecorresponding soft-typing formulation .…

Federated learning enables a cluster of decentralized mobile devices at theedge to collaboratively train a shared machine learning model, while keeping the raw training samples on device . This decentralized training approach isdemonstrated as a practical solution to mitigate the risk of privacy leakage .…

In first-orderlogic induction requires an infinite number of axioms, which is not a feasible input to a computer-aided theorem prover requiring a finite input . In this work we introduce a new method, inspired by the field of axiomatictheories of truth, that allows to express schematic inductive definitions .…

In this work, we study the $k$-median clustering problem with an additionalequal-size constraint on the clusters, from the perspective of parameterizedpreprocessing . Our main result is the first lossy ($2$-approximate) polynomial kernel for this problem, parameterized by the cost of clustering .…

In the Euclidean $k$-Means problem we are given a collection of $n$ points$D$ and a positive integer $k$. This problem is known to be APX-hard and the current best approximationratio is a primal-dual $6.357$ approximation . In this note we show how a minor modification of Ahmadian et al.’s…

CNF-based SAT and MaxSAT solvers are central to logic synthesis andverification systems . Increasing popularity of these constraint problems in electronic design automation encourages studies on different SAT problems . In this work, we propose a one-shot model derived from the Transformer architecture to solve theMaxSAT problem .…

Evolution-based neural architecture search requires high computational resources, resulting in long search time . In this work, we propose a framework of applying the Covariance Matrix Adaptation Evolution Strategy (CMA-ES) to theneural architecture search problem called CMANAS . We also used an architecture-fitness table (AF table) for keepingrecord of the already evaluated architecture, thus further reducing the searchtime .…

We present a technique that allows for improving on some relative greedy algorithms by well-chosen (non-oblivious) local search algorithms . Relativegreedy procedures are a particular type of greedy algorithm that start with asimple, though weak, solution, and iteratively replace parts of this startingsolution by stronger components .…

We introduce a new model for two-sided matching which allows us to borrow popular fairness notions from the fair division literature such asenvy-freeness up to one good and maximin share guarantee . We show that (a slight strengthening of) DEF1 cannot always be achieved, but in the special case where both sides have identical preferences, the round-robin algorithm with a carefully designed ordering achieves it .…

Randomizedsketching has emerged as a powerful technique for constructing estimates of the Hessian which can be used to perform approximate Newton steps . We show that the Gaussian sketchingmatrix can be drastically sparsified, significantly reducing the computational cost of sketching, without substantially affecting its convergence properties .…

In the Sparse Fast Fourier Transform (Sparse FFT) problem, one is given oracle access to a $d-dimensional vector $x$ of size $N$ and is asked to compute the best $k$-term approximation of its Discrete Fourier transform . This is in sharp contrast with the classical FFT algorithm of Cooley and Tukey, which is stable and completely insensitive to the dimension of the input vector .…

We consider alignment of sparse graphs, which consists in finding a mapping between nodes of two graphs which preserves most of the edges . We design an optimal test for this problem which givesrise to a message-passing algorithm for graph alignment .…

The study of fairness in multiwinner elections focuses on settings where candidates have attributes . Voters may also be divided into predefinedpopulations under one or more attributes . The models that focus on candidate attributes alone maysystematically under-represent smaller voter populations .…

NeuSaver reduces the power consumption of mobile devices when streaming videos by applying an adaptive frame rate to each video chunk without compromising user experience . The paper concludes that high-quality videos consume considerable amounts of energy on mobiledevices . It also uses anasynchronous advantage actor-critic algorithm to reinforce the RL model quickly and robustly.…

In this work, we introduce an optoelectronic spiking artificial neuroncapable of operating at ultrafast rates ($\approx$ 100 ps/optical spike) and with low energy consumption ($<$ pJ/spike) The proposed system combines anexcitable resonant tunnelling diode (RTD) element exhibiting negativedifferential conductance, coupled to a nanoscale light source (forming a masternode) or a photodetector . We also demonstrate an optically-interconnectedspiking neural network model for processing of spatiotemporal data at over 10Gbps with high inference accuracy . These results demonstrate thepotential and viability of RTD spiking nodes for low footprint, low energy, …

In this paper, we enriched Ant Colony Optimization (ACO) with intervaloutranking to develop a novel multiobjective ACO optimizer . This proposal is suitable if the preferences of the Decision Maker (DM) can be modeled through outranking relations . IO-ACO is the firstant-colony optimizer that embeds an outranking model to bear vagueness andill-definition of DM preferences .…

Recently, various deep neural networks have been applied to classifyelectroencephalogram (EEG) signal . EEG is a brain signal that can be acquired in a non-invasive way and has a high temporal resolution . The average accuracy of the proposed model was 77.70% for theBCI competition IV_2a data set.…

Image smoothing is a fundamental procedure in applications of both computervision and graphics . The inherent smoothing nature of one smoothing operator is usually fixed and thus cannot meet the requirements of different applications . A generalized framework is then proposed with the introduced truncated Huber penalty function .…

Sound synthesiser controls typically correspond to technical parameters of signal processing algorithms rather than intuitive sound descriptors . This makes it difficult to realise soundideas in a straightforward way . Cross-modal mappings, for example betweengestures and sound, have been suggested as a more intuitive control mechanism .…

Secure Computation (SC) is a family of cryptographic primitives for computing on encrypted data in single-party and multi-party settings . SC is beingincreasingly adopted by industry for a variety of applications . MAGE calculates the memory access pattern ahead of time and uses it to produce a memory management plan .…

Geometry processing of 3D objects is of primary interest in many areas of computer vision and graphics, including robot navigation, 3D objectrecognition, classification, feature extraction, etc. The aim of this thesis is topresent algorithms, mainly inspired by the spectral analysis, subspacetracking, etc, that can be used and facilitate low-level 3Dgeometry processing, pattern recognition tasks and high-level applications, such as registration and identification of 3 D objects in partially scanned and cluttered scenes, taking into consideration different 3D models such as static and dynamic point clouds, static anddynamic 3D meshes .…

In this work, we study a scenario where a publisher seeks to maximize itstotal revenue across two sales channels: guaranteed contracts that promise to deliver a certain number of impressions to advertisers . On the one hand, if a guaranteed contract is not fullydelivered, it incurs a penalty for the publisher .…

Secure Computation (SC) is a family of cryptographic primitives for computing on encrypted data in single-party and multi-party settings . SC is beingincreasingly adopted by industry for a variety of applications . MAGE calculates the memory access pattern ahead of time and uses it to produce a memory management plan .…

This paper introduces an optimization problem for proper scoring rule design . We propose an efficient algorithm to compute an optimal scoring rule when the collection of distributions is finite . We alsoprove the optimality of the log scoring rule over a smaller set of functionsfor categorical distributions with Dirichlet priors .…

In this paper we present the first deterministic near-linear-time algorithm for computing sparse nonnegative convolutions . This immediately gives improveddeterministic algorithms for the state-of-the-art of output-sensitive SubsetSum, block-mass pattern matching, $N-fold Boolean convolution, and others,matching up to log-factors the fastest known randomized algorithms for these problems .…

In this paper, a simple technique for Unmanned Aerial Vehicles (UAVs)potential landing site detection using terrain information throughidentification of flat areas, is presented . The algorithm utilizes digitalelevation models (DEM) that represent the height distribution of an area . Flatareas which constitute appropriate landing zones for UAVs in normal oremergency situations result by thresholding the image gradient magnitude of the digital surface model .…

Every representative democracy must specify a mechanism under which voters choose their representatives . The most common mechanism in the U.S. –winner-take-all single-member districts — both enables substantial partisangerrymandering and constrains `fair’ redistricting . We study the design of multi-memberdistricts (MMDs) in which each district elects multiple representatives, potentially through a non-winner-takes-all voting rule .…

In the Euclidean $k$-Means problem we are given a collection of $n$ points$D$ and a positive integer $k$. This problem is known to be APX-hard and the current best approximationratio is a primal-dual $6.357$ approximation . In this note we show how a minor modification of Ahmadian et al.’s…

In this paper, we propose a cross-modal variational auto-encoder (CMVAE) for content-based micro-video background music recommendation . CMVAE is ahierarchical Bayesian Bayesian generative model that matches relevant background musicto a micro video by projecting these two multimodal inputs into a shared low-dimensional latent space .…

Secure Computation (SC) is a family of cryptographic primitives for computing on encrypted data in single-party and multi-party settings . SC is beingincreasingly adopted by industry for a variety of applications . MAGE calculates the memory access pattern ahead of time and uses it to produce a memory management plan .…

Gale-Shapley introduced a matching problem between two sets of agents whereeach agent on one side has a preference over the agents of the other side . Shapley-Shubik, Demange-Gale, and many others extended the model by allowing monetary transfers . In thispaper, we study an extension where matched couples obtain their payoffs as the outcome of a strategic game .…