Transfer Learning for Human & AI

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 .…

A partial fix to a known software issue is a programmer’s failed attempt to fix the issue the first time . The result is abenchmark set of 2204 benchmark tasks for automated program repair based on partial fixes . The benchmark set is available open source and open to further contributions to further improvements and improvement .…

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 .…

ContrastiveDual Learning for Approximate Nearest Neighbor (DANCE) is an effective training paradigm for dense retrieval . DANCE incorporates an additionaldual training object of query retrieval, inspired by the classic informationretrieval training axiom, query likelihood . With contrastive learning, the dualtraining object of DANCE learns more tailored representations for queries and documents to keep the embedding space smooth and uniform, thriving on theranking performance of Dance on the MS MARCO document retrieval task .…

Alterations in the human gut bacteriome can be associated with human health issues, such as type-2 diabetes and cardiovascular disease . We propose atwo-layer network system to investigate its dynamics . We develop an in-silico simulation model (virtual GB) allowing us to study impact of metabolite exchange through molecular communications in the gutbacteriome network system .…

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.…

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 .…

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 .…

MetaCoq aims to provide the first fully-certified realistic implementation of a type checker for the full calculus underlying the Coq proofassistant . We show how theoretical tools such as bidirectional type-checking, Tait-Martin-L\”of/Takahashi’s confluence proof technique and monadic anddependently-typed programming can help construct the following artefacts: aspecification of Coq’s syntax and type theory, the Polymorphic Cumulative CumulativeCalculus of (Co)-Inductive Constructions (PCUIC); a monad for the manipulation of raw syntax and interaction with Coq system; a verification of PCUIC’smetatheory, whose main results are the confluence of reduction, typepreservation and principality of typing; a realistic, correct and completetype-checker for PCIC .…

Sampling is often a necessary evil to reduce the processing and storage cost of distributed tracing . The resulting traces are often onlypartially and not completely sampled which complicates statistical analysis . We describe a scalable and adaptivesampling approach that can preserve events of interest better than the widelyused head-based sampling approach .…

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 .…

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 .…

The set of integer number lists with finite length, and the set of binary trees with integer labels are countably infinite . Many inductively definedtypes also have countably many elements . In this paper, we formalize the syntaxof first order inductive definitions in Coq and prove them countable, undersome side conditions .…

Measuringinter-cluster reliability of multidimensional projection is crucial as it affects how well inter-luster tasks can be conducted . We propose Steadiness and Cohesiveness, two novel metrics to measure inter-clusters reliability . They extract random clusters with arbitrary shapes and positions in one space and evaluate how much the clusters are stretched or dispersed in the other space .…

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 .…

This paper presents an efficient and comprehensive MATLAB code to solve two-dimensional structural topology optimization problems . The Unsmooth Variational Topology Optimization (UNVARTOP) method, developed by the authors in a previous work, is used in the topologyoptimization code . The code, intended for students and newcomers intopology optimization, is included as an appendix (Appendix A) and it can bedownloaded from https://github.com/DanielYago…

The world is expecting an aging population and shortage of healthcareprofessionals . This poses the problem of providing a safe and dignified lifefor the elderly . Technological solutions involving cameras can contribute to safety, comfort and efficient emergency responses, but they are invasive ofprivacy .…

Online training materials and educational videos were more likely to identify misinformation than those in our controlgroup (total $N$=1000) Given the promising results of our preliminary study, we plan to expand efforts in this area, and build upon lessons learned from this pilot study .…

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 an efficient and comprehensive MATLAB code to solve two-dimensional structural topology optimization problems . The Unsmooth Variational Topology Optimization (UNVARTOP) method, developed by the authors in a previous work, is used in the topologyoptimization code . The code, intended for students and newcomers intopology optimization, is included as an appendix (Appendix A) and it can bedownloaded from https://github.com/DanielYago…

Zelig: our blockchainsimulator designed with the main goals of customizability and extensibility . Zelig is the only blockchain simulator that enablessimulating custom network topologies without modifying the simulator code . We explain our simulator design, validate via experimental analysis against thereal-world Bitcoin network, and highlight potential use cases .…

We study the $c-approximate near neighbor problem under the continuousFr\’echet distance . Given a set of $n$ polygonal curves with $m$ vertices, aradius $\delta 0$ and a parameter $k \leq m$ We obtain new upper bounds that providedifferent tradeoffs between approximation factor, preprocessing time, and querytime .…

This paper proposes a simple but effective data augmentation method . It uses pseudo-labelling to select samples from noisy distantsupervision datasets . The result shows that the proposed methodimproves the accuracy of biased news detection models . The proposed method shows that it can be used to improve news detection accuracy .…

Data minimization is a legal obligation defined in the European Union’s General Data Protection Regulation (GDPR) as the responsibility to process anadequate, relevant, and limited amount of personal data in relation to aprocessing purpose . However, unlike fairness or transparency, the principle has not seen wide adoption for machine learning systems due to a lack ofcomputational interpretation .…

There is a growing awareness of the importance of human values in software systems . We argue that a tool that can identify human values from softwaredevelopment artefacts can (partially) address this gap . Our study finds that a human values dashboard would be useful for the development team (e.g.,…

People park their vehicle depending on interior and exterior contexts, even unconsciously . Park4U Mate is a voice-based in-car assistant that is aware ofinterior and exterior context . The solution was demonstrated to thirty-five users in test-drives and their feedback was collected on the system’s decision-making capability as well as on the human-machine-interaction .…

The main challenges of Optical Music Recognition (OMR) come from the nature of written music, its complexity and the difficulty of finding an appropriatedata representation . DoReMi was generatedusing a music notation software and includes over 6400 printed sheet music images with accompanying metadata useful in OMR research .…

Deep reinforcement learning has generated superhuman AI in competitive gamessuch as Go and StarCraft . Can similar learning techniques create a superior AIteammate for human-machine collaborative games? Will humans prefer AI teammatesthat improve objective team performance or those that improve subjectivemetrics of trust?…

Consistent hasing has played a fundamental role as a data router and a loadbalancer in various fields, such as distributed database, cloud infrastructure, and peer-to-peer network . The existing consistent hashing schemescan’t meet the requirements simultaneously, including full consistency,scalability, small memory footprint, low update time and low query complexity .…

Loop-invariant synthesis is the basis of every program verificationprocedure . Despite the common belief that the design ofheuristics is vital for the effective performance of a verifier, little work has been performed toward obtaining the optimal heuristics for eachinvarant-synthesis tool .…

A partial fix to a known software issue is a programmer’s failed attempt to fix the issue the first time . The result is abenchmark set of 2204 benchmark tasks for automated program repair based on partial fixes . The benchmark set is available open source and open to further contributions to further improvements and improvement .…

Automated audio captioning (AAC) is the task of automatically creating descriptions for the contents of a general audiosignal . Most AAC methods are using existing datasets to optimize and/orevaluate upon . In this paper we present a first approach for continuouslyadapting an AAC method to new information, using a continual learning method .…

We present a system called Adelfa that provides mechanized support for reasoning about specifications developed in the Edinburgh Logical Framework orLF . Typing judgements in LF are represented by atomic formulas in L_LF and quantification is permitted overcontexts and terms that appear in such formulas .…

Statistical methods for reconstructing networks from repeated measurementstypically assume that all measurements are generated from the same underlying structure . We describe a Bayesian analysis framework for such data that allows for the fact that network measurements may be reflective of multiple possiblestructures .…

We study the $c-approximate near neighbor problem under the continuousFr\’echet distance . Given a set of $n$ polygonal curves with $m$ vertices, aradius $\delta 0$ and a parameter $k \leq m$ We obtain new upper bounds that providedifferent tradeoffs between approximation factor, preprocessing time, and querytime .…

Formal verification is the only form of sofware testing that can guarantee the absence of bugs . The implementation we propose is based on the LongMap of the Scala standard library with some minor adaptations . We give the specification withrespect to an implementation of a map based on a list of tuples, that we .plement…

Social networking and micro-blogging services, such as Twitter, play animportant role in sharing digital information . Currently, thereis no automated way of determining which news or users are credible and whichare not . Establishing a system that can measure the social media user’scredibility has become an issue of great importance .…

Social networking and micro-blogging services, such as Twitter, play animportant role in sharing digital information . Currently, thereis no automated way of determining which news or users are credible and whichare not . Establishing a system that can measure the social media user’scredibility has become an issue of great importance .…

Modern software development heavily uses versioned packages, but programming languages rarely support the concept of versions in their semantics . This paper proposes a programming language that intrinsically supports versions . The proposed core calculus, called Lambda VL, has versioned values, each containing different values under different versions .…

We study the $c-approximate near neighbor problem under the continuousFr\’echet distance . Given a set of $n$ polygonal curves with $m$ vertices, aradius $\delta 0$ and a parameter $k \leq m$ We obtain new upper bounds that providedifferent tradeoffs between approximation factor, preprocessing time, and querytime .…

This is the Proceedings of ICML 2021 Workshop on Theoretic Foundation,Criticism, and Application Trend of Explainable AI . Deep neural networks (DNNs) have brought great success to a wide range of applications incomputer vision, computational linguistics, and AI . Interpreting and theorizing the internalmechanisms of DNNs becomes a compelling yet controversial topic .…

Development of the intelligent autonomous robot technology presupposes its beneficial effect on the individuals and societies . In the case of disruptive emergent technology, not only questions of how to build, but also why to build and with what consequences are important .…

The minimization of makespan in open shop with time lags has been shownNP-hard in the strong sense even for the case of two machines and unit-time operations . The minimized makespan is NP-hard for weighted completion time or for jobs with release dates .…

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 .…

The main challenges of Optical Music Recognition (OMR) come from the nature of written music, its complexity and the difficulty of finding an appropriatedata representation . DoReMi was generatedusing a music notation software and includes over 6400 printed sheet music images with accompanying metadata useful in OMR research .…

Online training materials and educational videos were more likely to identify misinformation than those in our controlgroup (total $N$=1000) Given the promising results of our preliminary study, we plan to expand efforts in this area, and build upon lessons learned from this pilot study .…

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 .…

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 .…

We give a presentation of Pure type systems where contexts need not bewell-formed . We show that this presentation is equivalent to the usual one . The main motivation for this presentation was that, when we extend Pure type systemswith computation rules, we want to declare constants before the computation rules that are needed to check the type .…

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 .…