Transfer Learning for Human & AI

In this study, we propose a novel adversarial reprogramming (AR) approach for low-resource spoken command recognition . The AR procedure aims to modify the acoustic signals (from the target domain) to repurpose a pretrained SCR model . We evaluate the proposed AR-SCR system on three low- resource SCR datasets, including Arabic,Lithuanian, and dysarthric Mandarin speech .…

Deep language models such as BERT have achievedunprecedented success in many applications of computational linguistics . However, much less is known about how these models can be used to analyze existing text . We propose a novel method that combines transformer models withnetwork analysis to form a self-referential representation of language use within a corpus of interest .…

ACM “The Handbook on Socially Interactive Agents” reviews research on and technologies involving socially interactive agents . It outlines the design space for the appearance of agents, including what appearance comprises, modalities in which agents are presented, and how agents are constructed .…

A promising approach to reduce the backhaul load is toproactively cache content at the network edge, taking into account the overlaid social network . Known caching schemes require complete knowledge of the socialgraph and mainly focus on one-to-one interactions forgoing the prevalent mode of content sharing among circles of ‘friends’ We propose Bingo, a proactive content caching scheme that leverages the presence of interest groups in onlinesocial networks .…

Twitter is a social network that offers a rich and interesting source of information challenging to retrieve and analyze . The available operations allow retrieving tweets on the basis of a set of keywords but with limitations such as the number of calls perminute and the size of results .…

Memristor technologies have been rapidly maturing for the past decade to support emerging memory, artificial synapses, logic gates and bio-signal processing applications . Here, we report on the thermal properties of metal-oxide memristors and demonstrate how these can be used toemulate a fundamental function of biological neurons: homeostasis .…

We introduce the novel task of detecting sustainability initiatives incompany reports . Given a full report, the aim is to automatically identifymentions of practical activities that a company has performed in order totackle specific societal issues . As a single initiative can often be described over multiples sentences, new methods for identifying continuous sentence spans needs to be developed .…

I explain how faculty members could exploit a method to allocate travel funds . I also explain how to use game theory to design a method that cannot be manipulated .…

Pressure Field Contact (PFC) was recently introduced as a method for detailed modeling of contact interface regions at rates much faster thanelasticity-theory models . PFC was designed to work withcoarsely-meshed objects while preserving continuity to permit use of witherror-controlled integrators .…

Current research on Covid-19 is typically based on a single source of information, specifically on structured historical data . Other studies are exclusively focused on unstructured onlineretrieved insights, such as data available from social media . The proposed approach is based on vine copulas, which allow us to improve predictions byexploiting the dependencies between different sources of information .…

We study the problem of finding approximate first-order stationary points inoptimization problems . The objective function $f(x,y) issmooth, but assumed neither convex in $x$ nor concave in $y$ Our approachrelies upon replacing the function with its $k$th order Taylorapproximation (in $y) and finding a near-stationary point in the resultingsurrogate problem .…

The main purpose of Mining Software Repositories (MSR) is to discover the latest enhancements and provide an insight into how to make improvements in asoftware project . This paper updates the MSR findings of the original MSR Cookbook, by first conducting a systematic mapping study to elicitand analyze the state-of-the-art, and then proposing an extended version of the Cookbook .…

In Recommender Systems, more often than not, we design evaluations to measure an algorithm’sability to optimize goals in complex scenarios . We believe the RS community wouldgreatly benefit from creating a collection of standardized, simple, andtargeted experiments, which, much like a suite of “unit tests”, would assess algorithm’s ability to tackle core challenges that makeup complex RS tasks .…

The classical (parallel) black pebbling game is a useful abstraction which allows us to analyze the resources (space, space-time, cumulative space)necessary to evaluate a function $f$ with a static data-dependency graph $G$ . The pebbled complexity of thegraph $G  characterized the amortized cost of evaluating $f $f_{G,H$ multipletimes or the total cost to run a brute-force preimage attack over a fixed domain $mathcal{X}$ While a classical attacker will need to evaluate thefunction $f .…

Microservices Architecture (MSA) style is a promising design approach todevelop software applications consisting of multiple small and independently deployable services . We usedpeer-reviewed and grey literature to collect the patterns, strategies, and quality attributes for creating this decision model . The study proposes a decision model for selectingpatterns and strategies to decompose applications into microservices.…

This paper proposed a novel large neighborhood search-adaptive genetical algorithm (LNS-AGA) for many-to-many on-orbit repairing mission planning ofgeosynchronous orbit (GEO) satellites with mission deadline constraint . The missionobjective is to find the optimal servicing sequence and orbit rendezvous time of every servicing spacecraft to minimize total cost of all servicingspacecrafts with all target satellites repaired .…

Connection management is an important problem for any wireless network to ensure smooth and well-balanced operation throughout . Traditional methods for connection management (specifically user-cell association) consider sub-optimaland greedy solutions such as connection of each user to a cell with maximumreceive power .…

In decentralized machine learning, workers compute model updates on their local data . Because the workers only communicate with few neighbors without central coordination, these updates propagate progressively over the network . This paradigm enables distributed training on networks without all-to-allconnectivity, helping to protect data privacy as well as to reduce thecommunication cost of distributed training in data centers .…

Wi-Fi Protected Access3 (WPA3) has been shown to be vulnerable to several attacks . A deauthentication attack and a beacon flood attack, known to be possible on a WPA2 network, are still possible with WPA3 . We propose a design for asignature-based IDS, which incorporates techniques to detect all the aboveattacks .…

Social recommender systems are insufficient to distill the heterogeneous collaborative signals from both user and itemsides . We propose a Self-Supervised Metagraph Infor-max Network (SMIN) which investigates the potential of jointly incorporating social- andknowledge-aware relational structures into the user preference representation for recommendation .…

A directed acyclic graph $G=(V,E)$ is said to be $\epsilon$-depth robust if forevery subset $S \subseteq V$ of $|S| \leq e$ nodes . In cryptography, (extremely)depth-robust graphs with low indegree have found numerous applications including side-channel resistant Memory-Hard Functions, Proofs ofSpace and Replication .…

During the Coronavirus 2019 (the covid-19) pandemic, schools continuouslystrive to provide consistent education to their students . Teachers and education policymakers are seeking ways to re-open schools, as it is necessary for community and economic development . The School-Virus-Infection-Simulator (SVIS) simulates the spread of infection at a school .…

Knowledge-Enhanced Hierarchical Graph Transformer Network (KHGT) tackles multi-typed interactive patterns between users and items in recommendation systems . KHGT is built upon a graph-structured neuralarchitecture to capture type-specific behavior characteristics . Extensive experiments conducted on threereal-world datasets show that KHGT consistently outperforms many state-of-the-art recommendation methods .…

Social recommendation task aims to predict users’ preferences over items with the incorporation of social connections among users . KCGN enables the high-order user- and item-wise relation encoding byexploiting the mutual information for global graph structure awareness . We further augment KCGN with the capability of capturing dynamicmulti-typed user-item interactive patterns .…

Twitter is a social network that offers a rich and interesting source of information challenging to retrieve and analyze . The available operations allow retrieving tweets on the basis of a set of keywords but with limitations such as the number of calls perminute and the size of results .…

Digital imaging sensor technology has continued to outpace development inoptical technology in modern imaging systems . The resulting quality lossattributable to lateral chromatic aberration is becoming increasinglysignificant as sensor resolution increases . The goals ofhigher-performance and lighter lens systems drive a recent need to find new ways to overcome resulting image quality limitations .…

A new algorithm is proposed to solve the problem of partitioning a polygon into a set of connected disjoint sub-polygons, each of which covers an area of a specificsize . The work is motivated by terrain covering applications in robotics, where the goal is to find an efficient plans for a team of heterogeneous robotsto cover a given area .…

Big data has remarkably evolved over the last few years to realize anenormous volume of data generated from emerging services and applications . Federated learning (FL) appeared to be a promising learningtechnique . However, a gap exists in the literature that a comprehensive survey on FL for big data services is yet to be conducted .…

Energy forecasting has attracted enormous attention over the last few decades, with novel proposals related to the use of heterogeneous data sources,probabilistic forecasting, online learn-ing, etc. A key aspect that emerged isthat learning and forecasting may highly benefit from distributed data .…

We consider repeated first price auctions where each bidder, having adeterministic type, learns to bid using a mean-based learning algorithm . We characterize the Nash convergence property of the bidding dynamics in two senses: (1) time-average: the fraction of rounds where bidders play aNash equilibrium approaches to 1 in the limit; (2) last-iterate: the mixedstrategy profile of bidder approaches to a Nash equilibrium in limit .…

We consider repeated first price auctions where each bidder, having adeterministic type, learns to bid using a mean-based learning algorithm . We characterize the Nash convergence property of the bidding dynamics in two senses: (1) time-average: the fraction of rounds where bidders play aNash equilibrium approaches to 1 in the limit; (2) last-iterate: the mixedstrategy profile of bidder approaches to a Nash equilibrium in limit .…

The main interest of this paper is to study the relationship between the low-discrepancy sequence and the static solution to the multi-bodies problem in high-dimensional space . We also combine the DEM with therestarting technique to generate a series of low-Discrepancy sequences .…

Specialized accelerators provide gains of performance and efficiency inspecific domains of applications . Sparse data structures or/and representationsexist in a wide range of applications. PYXIS collects accelerator designs andreal execution performance statistics. Currently, there are 73.8 K instances in PYxIS. PyXIS is open-source, and we are constantly growing with new accelerator designs and performance statistics .…

In this position paper, we discuss recent applications of simulation approaches for recommender systems tasks . In particular, we describe how they were used to analyze the problem of misinformation spreading . We also present potential lines of future work wheresimulation methods could advance the work in the recommendation community .…

This paper introduces a model for producing stylized line drawings from 3Dshapes . The model takes a 3D shape and a viewpoint as input, and outputs adrawing with textured strokes, with variations in stroke thickness,deformation, and color learned from an artist’s style .…

Bonnet, Kim, Thomass\'{e, and Watrigant (2020) introduced the twin-width of a graph . We show that the twin width of an $n$-vertex graph is less than $(n+\sqrt{n\ln n}+\qrt {n}+2\lt n}$ asymptotically almost surely for any positive $varepsilon$ We also calculate the twinwidth of randomgraphs $G(n,p)$ with $p\leq c/n$ for a constant $c <1$ and $1$ to $2$ . …

Session-based recommendation plays a central role in a wide spectrum of online applications, ranging from e-commerce to online advertising services . We propose a multi-task learning framework with Multi-levelTransition Dynamics (MTD) which enables the jointly learning of intra- andinter-session item transition dynamics in automatic and hierarchical manner .…

We present a robust learning algorithm to detect and handle collisions in 3D deforming meshes . We use a numerical optimization algorithm to resolve penetrationsguided by the network . Our approach outperforms supervised learning methods and achieves$93.8-98.1\%$ accuracy compared to the groundtruth by analytic methods .…

Stereoscopic projection mapping (PM) allows a user to see a 3D computer-generated (3D) object floating over physical surfaces ofarbitrary shapes around us . The current technology only satisfies binocular cues and is not capable of providing correct focus cues, which causes a vergence–accommodation conflict(VAC) We propose a multifocal approach to mitigate VAC instereoscopic PM .…

We consider repeated first price auctions where each bidder, having adeterministic type, learns to bid using a mean-based learning algorithm . We characterize the Nash convergence property of the bidding dynamics in two senses: (1) time-average: the fraction of rounds where bidders play aNash equilibrium approaches to 1 in the limit; (2) last-iterate: the mixedstrategy profile of bidder approaches to a Nash equilibrium in limit .…

During the Coronavirus 2019 (the covid-19) pandemic, schools continuouslystrive to provide consistent education to their students . Teachers and education policymakers are seeking ways to re-open schools, as it is necessary for community and economic development . The School-Virus-Infection-Simulator (SVIS) simulates the spread of infection at a school .…

The field of quantitative analytics has transformed the world of sports over the last decade . We often view teamsports, such as soccer, hockey, and baseball, as pairwise win-lose encounters . We propose a novel model of this encounter as a zero-sum stochastic game, in which the goal of the batter is toget on base, an outcome the pitcher aims to prevent .…

Timeout bugs can cause serious availability and performance issues which are difficult to fix due to the lack of diagnostic information . In this paper, we present TFix+ a self-configuring timeout bug fixingframework for automatically correcting two major kinds of timeout bugs with dynamic timeout valuepredictions .…

Aspect-Oriented Requirements Engineering (AORE) extends the existing software engineering approaches to cope with the issue of tangling andscattering resulted from crosscutting concerns . The proposed framework comprises a high-level and low-levelvalidation to implement on software requirements specification (SRS) Validation of requirements artefacts is an essential task in software development .…

Neural Architecture Search (NAS) methodology is becoming increasingly important for both academia and industries . Most existing NAS methods are computationally expensive for real-world deployments . We introduce a new estimation metric, named Random-Weight Evaluation (RWE) to quantify the quality of CNNs in a cost-efficient manner .…

Pressure Field Contact (PFC) was recently introduced as a method for detailed modeling of contact interface regions at rates much faster thanelasticity-theory models . PFC was designed to work withcoarsely-meshed objects while preserving continuity to permit use of witherror-controlled integrators .…

Network slicing is a promising technology that allows mobile network operators to efficiently serve various emerging use cases in 5G . It is challenging to optimize the utilization of network infrastructures while ensuring the performance of network slices according to service levelagreements (SLAs) To solve this problem, we propose SafeSlicing that introduces a new constraint-aware deep reinforcement learning (CaDRL) algorithm to learn the optimal resource orchestration policy within two steps, i.e.,offline…

Intrinsic projector calibration is essential in projection mapping (PM) applications, especially in dynamic PM . We aim to estimate the intrinsic parameters of a projector while avoiding the limitation of shallow DOF . The device consists of a flat-bed scanner and pinhole-array masks .…

Modern recommender systems often embed users and items into low-dimensionallatent representations, based on their observed interactions . Exploring multi-typed behavior patterns is of great importance torecommendation systems, yet is very challenging because of two aspects: i. The complex dependencies across different types of user-item interactions; ii)Diversity of such multi-behavior patterns may vary by users due to their personalized preference .…

BOOST is a novel self-supervised model to focus pre-training based on the characteristics of source code . We train our model in a way that brings the functionally equivalent code closer and distinct code further through a contrastive learning objective .…