On the Hardness of Compressing Weights

We investigate computational problems involving large weights through the lens of kernelization . Our main focus is the weighted Cliqueproblem, where we are given an edge-weighted graph and the goal is to detect aclique of total weight equal to a prescribed value .…

Accessible Color Cycles for Data Visualization

Data were collected with an online survey, and the results were used to train a machine-learning model . Optimal color cycles containing six, eight, and tencolors were generated using the data-driven aesthetic-preference model and accessibility constraints . Due to the balance of aesthetics and accessibility considerations, the resulting color cycles can serve as reasonable defaults indata-plotting codes for data visualization .…

Decision problems for origin close top down tree transducers full version

Tree transductions are binary relations of finite trees . For treetransductions defined by non-deterministic top-down tree transducers,inclusion, equivalence and synthesis problems are known to be undecidable . We introduce a similarity measure for transducers with origin semantics . We show that we can decide inclusion and equivalence problems for origin-close non-dualistic top-downs tree transducers .…

Analyzing a Knowledge Graph of Industry4 0 Standards

In this article, we tackle the problem of standard interoperability across different standardization frameworks, and devise a knowledge-driven approach . The STO ontology representsproperties of standards and standardization . The I40KG integrates more than 200 standards and fourstandardization frameworks . We analyze both thenumber of discovered relations between standards and accuracy of theserelations .…

A Serverless Cloud Integration For Quantum Computing

The proposed solution uses the up-to-date cloud native technologies (e.g. CloudFunctions, Containers, Microservices) and serves as a general framework todevelop multiple applications on the same infrastructure . Every triggeredrun leverages on multiple backend options, that include a scheduler managing the queuing mechanism to correctly schedule jobs and final results retrieval .…

ATC an Advanced Tucker Compression library for multidimensional data

ATC is a C++ library for advanced Tucker-based compression of numerical data . It is based on the sequentially truncated higher-order singular value decomposition (ST-HOSVD) and bit plane truncation . Numerical results show that ATC maintains state-of-the-art Tucker compression rates, while providing average speed-ups of 2.6-3.6 and halving memory usage .…

Recovering the Unbiased Scene Graphs from the Biased Ones

Scene graph generation (SGG) aims to producecomprehensive, graphical representations describing visual relationships among objects . However, the imbalance in the fraction of missing labels of differentclasses, or reporting bias, exacerbating the long tail is rarely considered and cannot be solved by the existing debiasing methods .…

TransformerFusion Monocular RGB Scene Reconstruction using Transformers

TransformerFusion is a transformer-based 3D scene reconstruction approach . From an input monocular RGB video, the video frames are processed by a transformer network that fuses the observations into a volumetric featuregrid representing the scene . This feature grid is then decoded to a higher-resolutionscene reconstruction, using an MLP-based surface occupancy prediction frominterpolated coarse-to-fine 3D features .…

Repulsive Surfaces

This paper develops a numerical framework for optimization of surface geometry while avoiding (self-)collision . The starting point is thetangent-point energy, which effectively pushes apart pairs of points that are close in space but distant along the surface . We develop a discretization ofthis energy for triangle meshes, and introduce a novel acceleration scheme .…

A Data Driven Method for Recognizing Automated Negotiation Strategies

Understanding an opponent agent helps in negotiating with it . Wepropose a novel data-driven approach for recognizing an opponent’s snegotiation strategy . The approach includes a data generation method for anagent to generate domain-independent sequences by negotiating with a variety of opponents across domains, a feature engineering method for representingnegotiation data as time series with time-step features and overall features, and a hybrid (recurrent neural network-based) deep learning method .…

State efficient QFA Algorithm for Quantum Computers

Moore-Crutchfield quantum finite automaton(MCQFA) is proven to be exponentially more succinct than classical finiteautomata models . In this paper, we present a modified MCQFA algorithm for the language $mathtt{MOD}_{\rm p}$, the operators of which are selected based on the basis gates on the availablereal quantum computers .…

A Lottery Ticket Hypothesis Framework for Low Complexity Device Robust Neural Acoustic Scene Classification

We propose a novel neural model compression strategy combining dataaugmentation, knowledge transfer, pruning, and quantization for device-robustacoustic scene classification . Acoustic Lottery could compress an ASC model over$1/10^{4}$ and attain a superior performance (validation accuracy of 74.01% and log loss of 0.76) compared to its not compressed seed model .…

Demonstration of Faceted Search on Scholarly Knowledge Graphs

Traditional scholarly search systems listdocuments instead of providing direct answers to the search queries . As data in knowledge graphs are not acquainted semantically, they are not machine-readable . Instead, a search on scholarly knowledge graphs ends up in a full-text search, not a search in the content of scholarly literature .…

Dependent Type Systems as Macros

Turnstile+ is a high-level, macros-based metaDSL for building dependently typed languages . With it, programmers may rapidly prototype anditerate on the design of new dependently typing features and extensions . The framework’s support of language-oriented programming alsomakes it suitable for experimenting with systems of interacting components,e.g.,…

Persistent Cup Length

In this paper, we lift the cup-length into the Persistent Cup-Lengthinvariant for the purpose of extracting non-trivial information about the evolution of the cohomology ring structure across a filtration . We show that the Cup-length Diagram can be computed from a family of representativecocycles and devise a polynomial time algorithm for the computation of thePersistent cup-Length .…

The Semantics of Package Management via Event Structures

We propose an approach to the semantics of package management which relatesit to general event structures, well-known mathematical objects used in these mathematical objects . In this approach, the data of a package repository is treated as a declarative specification of anondeterministic, concurrent program .…

Instagrammable Data Using Visuals to Showcase More Than Numbers on AJ Labs Instagram Page

Use of data-driven storytelling is becoming increasingly integrated into theever-complex business models of news outlets, generating more impact and visibility . Previous studies have focused on studying these two effects separately . This paper identifies andanalyzes the use of data journalism on the Instagram content of AJ Labs, the team dedicated to producing data- driven and interactive stories for the AlJazeera news network .…

Certifying DFA Bounds for Recognition and Separation

Wesuggests to use determinacy of turn-based two-player games to generate certificates for the number of states that adeterministic finite automaton (DFA) needs in order to recognize a given language . We show that our approach is useful also for certification of separation of regular languages by a DFA of a given size .…

Linear Time Model Checking Branching Processes

Multi-type branching processes are a natural and well-studied model forgenerating random infinite trees . Branching processes feature bothnondeterministic and probabilistic branching, generalizing both transitionsystems and Markov chains . Westudy the complexity of model checking branching processes against linear-timeomega-regular specifications . The underlying general model-checking algorithm is based on the automata-theoretic approach, using unambiguous B\”uchi automata .…

Traffic Signal Control with Communicative Deep Reinforcement Learning Agents a Case Study

Two recently proposed multi-agent reinforcement learning methods can be applied to control traffic signals in urban areas . The two methods differ intheir use of a reward calculated locally or globally and in the management of agents’ communication . MA2C achieves the best performance in themajority of cases, outperforms the alternative method considered, and displays stability during the learning process .…

Hierarchical Planning for Dynamic Resource Allocation in Smart and Connected Communities

Resource allocation under uncertainty is a classical problem in city-scalecyber-physical systems . Consider emergency response as an example; urbanplanners and first responders optimize the location of ambulances to minimize response times to incidents such as road accidents . We present a general approach to hierarchical planning that leverages structure in city level CPS problems for resource allocation .…

Trans4E Link Prediction on Scholarly Knowledge Graphs

The incomplepleteness of Knowledge Graphs (KGs) is a crucial issue affecting the quality of AI-based services . In the scholarly domain, KGs describing researchpublications typically lack important information, hindering our ability to analyze and predict research dynamics . In this work, we present Trans4E, a novel embedding model that isparticularly fit for KGs which include N to M relations with N$\gg$M .…

Audio visual Attentive Fusion for Continuous Emotion Recognition

We propose an audio-visual spatial-temporal deep neural network with: (1) avisual block containing a pretrained 2D-CNN followed by a temporalconvolutional network . (2) an aural block containing several parallelTCNs; and (3) a leader-follower attentive fusion block combining theaudio-visual information . The fusion block emphasizes the visual modality while exploits thenoisy aural modality using the inter-modality attention mechanism .…

Solving the subset sum problem with a nonideal biological computer

We consider the solution of the subset sum problem based on a parallelcomputer consisting of self-propelled biological agents moving in ananostructured network that encodes the NP-complete task in its geometry . Wedeveloped an approximate analytical method to analyze the effects of small errors in the nonideal junctions composing the computing network by using a Gaussianconfidence interval approximation of the multinomial distribution .…

Robust multigrid techniques for augmented Lagrangian preconditioning of incompressible Stokes equations with extreme viscosity variations

We present augmented Lagrangian Schur complement preconditioners and robustmultigrid methods for incompressible Stokes problems with extreme viscosityvariations . Such Stokes systems arise, for instance, upon linearization of nonlinear viscous flow problems . We extend a multigrid schemewith a discretization-dependent smoother and transfer operators fromtriangular/tetrahedral to quadrilateral/hexahedral finite elementdiscretizations .…

Hafnia based Double Layer Ferroelectric Tunnel Junctions as Artificial Synapses for Neuromorphic Computing

Ferroelectric tunnel junctions (FTJ) based on hafnium zirconium oxide(Hf1-xZrxO2) are a promising candidate for future applications, such aslow-power memories and neuromorphic computing . The tunneling electroresistance(TER) is tunable through the polarization state of the HZO film . By exploiting previously reported switching behaviour and thegradual tunability of the resistance, FTJs can be used as potential candidates for the emulation of synapses for neuromorphic .…

Instagrammable Data Using Visuals to Showcase More Than Numbers on AJ Labs Instagram Page

Use of data-driven storytelling is becoming increasingly integrated into theever-complex business models of news outlets, generating more impact and visibility . Previous studies have focused on studying these two effects separately . This paper identifies andanalyzes the use of data journalism on the Instagram content of AJ Labs, the team dedicated to producing data- driven and interactive stories for the AlJazeera news network .…

Web Archive Analytics

Web archive analytics is the exploitation of publicly accessible web pages and their evolution for research purposes . The Webis research group, a network of university chairs to which the authors belong, concluded an agreement with the Internet Archive todownload a substantial part of its web archive for research .…

Structural biology in the clouds The WeNMR EOSC Ecosystem

Structural biology aims at characterizing the structural and dynamicproperties of biological macromolecules at atomic details . Gaining insight intothree dimensional structures of biomolecules and their interactions is critical for understanding the vast majority of cellular processes . Thematic Services in the European Open Science Cloud (EOSC) portal(www.eosc-portal.eu)…

CLIP It Language Guided Video Summarization

A generic video summary is an abridged version of a video that conveys thewhole story and features the most important scenes . Users should have the option ofcustomizing the summary by using natural language to specify what is important . We propose a language-guidedmultimodal transformer that learns to score frames in a video based on their importance relative to one another and their correlation with a user-defined query .…

Chalcogenide optomemristors for multi factor neuromorphic computation

Nanosized films ofchalcogenide semiconductors can serve as building-blocks for novel types of computations where their electronic and optical properties are jointly exploited . We demonstrate that ultrathin photoactive cavities ofGe-doped Selenide can emulate the computationally powerful non-linearoperations of neo-Hebbian plasticity and the shunting inhibition .…