Transfer Learning for Human & AI

Supporting Clustering with ContrastiveLearning (SCCL) is a novel framework to leverage contrastive learning topromote better separation . SCCL significantly advances the state-of-the-art results on most benchmark datasets with 3%-11% improvement on Accuracy and 4%-15% improvements on Normalized Mutual Information . The framework leverages the strengths of both bottom-up instance discrimination and top-down clustering to achieve better intra-clusters and inter-cluster distances when evaluated with ground truth cluster labels .…

The lack of comprehensive sources of accurate vulnerability data represents acritical obstacle to studying and understanding software vulnerabilities (andtheir corrections) Inconclusion, our method reduces considerably the effort needed to search OSSrepositories for the commits that fix known vulnerabilities . We evaluated our approach using a prototype implementation named Prospector on a manually curated data set that comprises 2,391 known fix commits .…

This paper introduces a publicly available U.S.Broadband Coverage data set that reports broadband coverage percentages at azip code-level . We describe our error rangecalculation method and show that this additional data metric does not induceany privacy losses . We also explain how we used differential privacy to guaranteethat the privacy of individual households is preserved.…

Low-bit quantization of network weights and activations can drastically reduce the memory footprint, complexity, energy consumption and latency of DeepNeural Networks (DNNs) However, it can also cause an aconsiderable drop in accuracy when applied to complex tasks or lightweight DNNs .…

Living soft tissues appear to promote the development and maintenance of apreferred mechanical state within a defined tolerance around a so-called set-point . This phenomenon is often referred to as mechanical homeostasis . We present a novel finiteelement based computational framework that is constructed bottom up .…

Computer security research now takes into account the human factor by studying regular users and experts like operators and developers of systems . We analyzed the past decade of human factorsresearch in security and privacy, identifying 557 relevant publications . We found 48 publications focused on expert users and analyzed all indepth.…

Physical unclonable functions are promising candidates as security primitives for resource-constrained IoT devices . Arbiter PUFs (APUFs) are a group of delay-based PUFs which are highly lightweight in resource requirements but suffer from high susceptibility to machine learning attacks . The challenge input interface incurs low resource overhead and improves PUFs’ resistance against machine-learning attacks .…

Detecting piano pedalling techniques in polyphonic music remains achallenging task in music information retrieval . In this paper, we propose atransfer learning approach for the detection of sustain-pedal techniques . In the target task, the knowledge learned from the synthesised data can be transferred to detect the sustain pedal in acoustic piano recordings .…

Meta-ViterbiNet is a DNN-aided symbol detector that tracks channel variations with reduced overhead . It outperforms the previous best approach by a margin of up to 0.6dB in bit error rate in various challenging scenarios . It is based on a model-based/data-driven equalizer that operates with a relatively small number of trainable parameters .…

The theory of learning in games has extensively studied situations where agents respond dynamically to each other in light of a fixed utility function . However, in many settings of interest, agent utility functions themselves vary as a result of past agent choices .…

Two-dimensional thin plates are widely used in many applications . Shuntdamping is a promising way for the attenuation of vibration of theseelectromechanical systems . It enables a compact vibration damping method without adding significant mass and volumetric occupancy . This paper presents a methodology and aformulation for separately shunted piezoelectric patches for achieving higherperformance on vibration attenuation .…

The central challenge is how to securelydetect software exploits with minimal overhead . DIALED works in tandem with a companion CFA scheme to detect all (currently known) types of runtime software exploits at low-cost . Such attacks exploit softwarevulnerabilities to corrupt intermediate computation results stored in datamemory, changing neither the program code nor its control flow .…

Big languages such as English and Finnish have many natural languageprocessing (NLP) resources and models, but this is not the case for low-resourced and endangered languages as such resources are so scarce . In this paper, we present a method for constructing word embeddings for endangered languages .…

Offline delegation offers an efficient way to exchange coins . However, in such an approach, the coins thathave been delegated confront the risk of being spent twice . In DelegaCoin, an owner can delegate his coinsthrough offline-transactions without interacting with the blockchain network .…

Topic Modeling is used to analyze a collection of French Drama of the Classical Age and the Enlightenment . The aim of this contribution is to discover what semantic types of topics are found in this collection . This contribution showsthat interesting topic patterns can be detected which provide new insights into the structure of French drama as well as into the history of the classical Age and Enlightenment.…

The field of mathematical morphology offers well-studied techniques for imageprocessing . We demonstrate that morphological operations naturally form amultiparameter filtration and that persistent homology can then be used to extract information about both topology and geometry in the images . For illustration, we apply this framework to analyze noisy binary,grayscale, and color images.…

Cyber attacks are becoming more frequent and sophisticated . Organizations need to collaborate and share cyber threat information (CTI) and use it to improve their security posture . In this paper, we present TRADE — TRusted AnonymousData Exchange — a collaborative, distributed, trusted, and anonymized CTI sharing platform based on blockchain technology .…

This letter analyzes a class of information freshness metrics for large IoT systems in which terminals employ slotted ALOHA to access a common channel . The penalty function that follows a power law of the time elapsed sincethe last received update, generalizing the age of information metric .…

In this work, we study the problem of user preference learning on the exampleof parameter setting for a hearing aid . We propose to use an agent thatinteracts with a HA user, in order to collect the most informative data, andlearns user preferences for HA parameter settings .…

Auto-bidding has become one of the main options for bidding in online advertisements . Advertisers only need to specify high-level objectives and leave the task of bidding to auto-bidders . In this paper, we propose a family of auctions with boosts to improve welfare in auto-bid environments with both return on ad spend constraints and budget constraints .…

Virtual-Cluster is a new multi-tenant framework that extends Kubernetes . VirtualCluster provides both control plane and data plane isolations while sharing the underlying compute resources among tenants . The new framework preserves API compatibility by avoiding modifying the KuberNETes core components .…

Remote Direct Memory Access(RDMA) NICs (RNICs) are one such device, allowing applications to offloadremote memory accesses . RDMA still requires CPU intervention for complex offloads, beyond simple remote memory access . RedN can outperform one and two-sided RDMA implementations by up to 3x and 7.8x for key-value get operations and .performance…

The paper considers a distributed version of deep reinforcement learning(DRL) for multi-agent decision-making process in the paradigm of federatedlearning . Since the deep neural network models in federated learning are trained locally and aggregated iteratively through a central server, frequentinformation exchange incurs a large amount of communication overheads .…

We introduce and analyze a space-time hybridized discontinuous Galerkin method for the evolutionary Navier–Stokes equations . Key features of thenumerical scheme include point-wise mass conservation, energy stability, and pressure robustness . We prove that there exists a solution to the resulting nonlinear algebraic system in two and three spatial dimensions .…

Multi-column dependencies in relational databases come associated with twodifferent computational tasks . The detection problem is to decide whether adependency of a certain type and size holds in a given database . The discovery of inclusion-wise minimal UCCs is proven to be equivalent to the transversal hypergraph problem of enumerating the minimal hitting sets of a hypergraph .…

We conduct an empirical study of unsupervised neural machine translation for truly low resource languages . We show how adding comparable data mined using a bilingual dictionary along with modestadditional compute resource to train the model can significantly improve its performance .…

GitHub introduced GitHub Actions, a feature providing automated workflows for repository maintainers . Our research is the first to investigate how developers use Actions and how several activityindicators change after their adoption . Our results indicate that although only a small subset of repositories adopted GitHub Actions to date, there is a positive perception of the technology .…

DELFT builds a free-text knowledge graph fromWikipedia, with entities as nodes and sentences in which entities co-occur asedges . For each question, DELFT finds the subgraph linking question entitynodes to candidates using text sentences as edges . A novel graph neural network reasons over thefree-text graph-combining evidence on the nodes via information along edgesentences-to select a final answer .…

In this paper we investigate the problem of controlling a partially observed dynamical system such that its state is difficult to infer using a(fixed-interval) Bayesian smoother . This problem arises naturally in applications in which it is desirable to keep the entire state trajectory of asystem concealed .…

Reconfigurable intelligent surface (RIS) is a promising reflective radiotechnology for improving the coverage and rate of future wireless systems . The current work mainlyfocuses on the physical layer design of RIS . But enabling multiple devicesto communicate with the assistance of RIS is a crucial challenging problem .…

We present a type system for strategy languages that express programtransformations as compositions of rewrite rules . Our row-polymorphic typesystem assists compiler engineers to write correct strategies by . statically rejecting compositions of rewrites that otherwise would . otherwise would failduring rewriting at runtime .…

Process discovery methods have obtained remarkable achievements in ProcessMining, delivering comprehensible process models to enhance managementcapabilities . But selecting the suitable method for a specific event loghighly relies on human expertise, hindering its broad application . Solutionsbased on Meta-learning (MtL) have been promising for creating systems with reduced human assistance .…

The study examined the participation of female students of South EasternNigerian tertiary institutions in Information and Communication Technologies(ICTs) The study discussed the attendant gender divide in ICTs participation . Findings suggest that high cost of ICT and highlevel of male dominance, which made females think that ICT is for males were the major reasons for low female participation .…

Compiler optimization is the most common source contributing to binarycode differences in syntax . We tailor search-basediterative compilation for the auto-tuning of binary code differences . We run BinTuner with GCC 10.2and LLVM 11.0 on SPEC benchmarks (CPU2006 & CPU2017), Coreutils, and OpenSSL .…

Little work has been conducted on how often these variables are documented in code comments and what types of knowledge the comments provide about variables of primitive types . Primitive types are fundamental components available in any programminglanguage . We present an approach fordetecting primitive variables and their description in comments using lexicalmatching and advanced matching .…

We consider a multi-stage facility reallocation problems on the real line, where a facility is being moved between time stages based on the locations of $n$ agents . Westudy this problem both in the offline setting and online setting . In the online setting the algorithm does not know these future locations and must decide the location of the facility on astage-per-stage basis .…

Atomic Switch Networks (ASN) comprising silver iodide (AgI) junctions, amaterial previously unexplored as functional memristive elements within highly-interconnected nanowire networks, were employed as a neuromorphicsubstrate for physical Reservoir Computing . This new class of ASN-based devices has been physically characterized and utilized to classify spoken digitaudio data, demonstrating the utility of substrate-based device architectures .…

A perspective paper will revolve around the role of radio localization in all the cellular generations, from firstgeneration to 6G . It speculates about the idea of holographic localizationenabled by the advent of large intelligent surfaces, made of metamaterials . The road ahead is the road ahead with the inherent challenges brought by this new technology, the paper says .…

The search for energy-efficient alternatives for computer data processingurges us to look at the neural architecture of our brain . Webuilt a neural network in which the magnetization of a multilayered CoPt sampleas an artificial synapse adapts to ultrashort optical pulses .…

The proliferation of social media platforms like Twitter has heightened theconsequences of the spread of misinformation . To understand and model thespread of misinformation, in this paper, we leveraged the SEIZ (Susceptible,Exposed, Infected, Skeptics) epidemiological model . Applying a mathematical model can help to understand the trends and dynamics of spread ofmisinformation on Twitter and ultimately help to develop techniques to quicklyidentify and control it .…

We obtain the first tight continuityestimate on the Shannon entropy of random variables with a countably infinitealphabet . The proof relies on a new mean-constrained Fano-type inequality and the notion of maximal coupling . The above scheme works only for Shannon- and von Neumann entropies .…

This paper describes an approach to solving the next destination city recommendation problem for a travel reservation system . We propose a two stagesapproach: a heuristic approach for candidates selection and an attention neuralnetwork model for candidates re-ranking . We used this approach to solve the Booking.com…

We develop the learning algorithm to build the architecture agnostic model of the reconfigurable optical interferometer . Programming the unitarytransformation on the optical modes of the interferometers either follows theanalytical expression or requires the optimization routine if the analytic decomposition does not exist .…

Neural keyphrase generation models have recently attracted interest in their ability to output absent keyphrases . In this paper, we discuss the usefulness of absentkeyphrase from an Information Retrieval (IR) perspective . We introduce a finer-grained categorization scheme that sheds light on the impact on scientific document retrieval .…

In this paper, we study linear filters to process signals defined onsimplicial complexes . We propose a finite impulse response filter based on the HodgeLaplacian . We demonstrate how this filter can be designed to amplify orattenuate certain spectral components of simplicial signals .…

Understanding how beliefs are accepted, spread, and intensified is critical for any attempt to mitigate radicalization and avoid increased political polarization . This article presents an agent-based model of a social media network that enables investigation of the effects of censorship on the amount of dissenting information to which agents become exposed and thecertainty of their radicalized views .…

Money laundering (ML) is the behavior to conceal the source of money achievedby illegitimate activities . CubeFlow is a scalable, flow-based approach to spot fraud from a mass of transactions by modeling them as two coupled tensors and applying a novelmulti-attribute metric which can reveal the transfer chains accurately .…

Contact tracing via digital tracking applications installed on mobile phones is an important tool for controlling epidemic spreading . Our results are based on a combination of explicitsimulations and mean-field analysis . They indicate that there can be majordifferences in the epidemic size and epidemic probabilities which are equivalent in the normal SIR processes.…

TeCoMiner is not based on some generative probabilistic model but ontopological considerations about co-occurrence networks of terms . We outlinethe methods used for identifying topics, describe the features of the tool, andsketch an application, using a corpus of policy related scientific news on environmental issues published by the European Commission over the last decade .…