Transfer Learning for Human & AI

This paper describes a series of projects that explore the possibilities ofmusical expression through the combination of pre-composed, interlocking,modular components . In particular, this paper presents a modular soundtrack recently composed by the author for “Currents of Creativity,” a permanentinteractive video wall installation at the Pope John Paul II Cultural Center .…

MATRIX (Multipurpose Array of Tactile Rods for Interactive eXpression) is a new musical interface for amateurs and professionals alike . It gives users a3-dimensional tangible interface to control music using their hands . It can be used in conjunction with a traditional musical instrument and a microphone, or as a stand-alone gestural input device .…

The focus of our effort is to provideuser-friendly and enjoyable experiences for new music and sound performances . Multimodality of musical instruments is explored in each attempt . The degree of controlability in the performance and the richness of expressions are alsodiscussed for each installation .…

The technologies behind passive resonant magnetically coupled tags are presented . Their application as a musical controller is illustrated by group performances, interactive installations and music toys .…

Accordiatron is a new MIDI controller for real-time performance based on the paradigm of a conventional squeeze box or concertina . It translates thegestures of a performer to the standard communication protocol of MIDI . When used in conjunction with a realtime signal processing environment, theAccordiatron becomes an expressive, versatile musical instrument .…

We present an acoustic musical instrument played through a resonance model of another sound . The resonance model is controlled in real time as part of thecomposite instrument . Our implementation uses an electric violin whose spatialposition modifies filter parameters of the resonance model .…

This paper reviews the existing literature on input device evaluation and evaluation in human-computer interaction (HCI) It discusses possible application of this knowledge to the design and evaluation of new interfaces for musical expressions . Specifically, a set of musical tasks is suggested to allow the evaluation of different existing controllers .…

In this paper we describe our efforts towards the development of liveperformance computer-based musical instrumentation . Our design criteria include ease of use coupled with a long term potential for virtuosity, minimaland low variance latency, and clear and simple strategies for programming therelationship between gesture and musical result .…

Over the last four years, we have developed a series of lectures, labs and assignments aimed at introducing enough technology so that students can design and build innovative interface devices .…

Most of the work in facialaction processing has been directed at automatically recognizing affect from facial actions . This paper reviews several projects on vision-based interfaces that rely on facialaction for intentional HCI . Applications to several domains are introduced, including text entry, artistic and musical expression and assistive technology for motor-impaired users .…

We present an analysis of literature relating to the evaluationmethodologies of Digital Musical Instruments . We then apply choice aspects from these evaluation models and apply them to an optimized evaluation for new DMIs. We then applied choice aspects of these models to an evaluation model .…

Real world applications of image retrieval such as Adobe Stock need a way for users to find images that are both visually (i.e. aesthetically) and conceptually (i .e. containing the same salient objects) as a query image . We learn a joint vision andconcept embedding in the same high-dimensional space .…

There has been an increasingly popular trend in Universities for curriculumtransformation to make teaching more interactive and suitable for onlinecourses . When questions are asked in a lecture there is a negligible chance of having similar questions repeatedly, but asynchronously this is more likely .…

We analyze Twitter data relating to the COVID-19 pandemic using dynamic topicmodeling techniques to learn topics and their prevalence over time . Topics arelearned using four methods: nonnegative matrix factorization (NMF), nonnegativeCP tensor decomposition (NCPD), online NMF, and online NCPD .…

The matching function used by most graph-based representation learning methods is the inner product, this linear operation can not fit complex semantics well . In thispaper, we introduce easily accessible textual information to alleviate the negative effects of data sparsity .…

A positive coronavirus patient can be traced through CDRA and contact tracing . The technique can track the path traversed by the patient and collect the cellnumbers of all those people who have met with the patient . The generalpublic is informed in advance to avoid the path followed by the patients .…

User modeling is critical for many personalized web services . Many existing methods cannot exploit useful information in unlabeled user behavior data . We propose two self-supervision tasks for user model pre-training . Thepre-trained user models are finetuned in downstream tasks to learntask-specific user representations.…

SumGNN outperforms the best baseline by up to 5.54\%, and performance gain is particularly significant in low data relationtypes . SumGnn:~~{\it knowledgesummarization graph neural network. In addition to a subgraph extraction module that can efficiently anchor on relevant subgraphs from a KG, a self-attention based subgraph summarization scheme to generate reasoning path within the subgraph, and a multi-channel knowledge and data integration module.that…

In this paper, we study the mean square error (MSE) and the bit error rate (BER) performance of the box-relaxation decoder in massivemultiple-input-multiple-output (MIMO) systems . Our analysis assumes that the number of transmit and receive antennas ($n$and$m$) grow simultaneously large while their ratio remains fixed .…

Far-field Wireless Power Transfer (WPT) has emerged as a viable power source for Internet of Things (IoT) and Wireless Sensor Network (WSN) Expansion of the power transfer range is an important challenge to enable the design of efficient networks of small autonomous devices and drive the massive numbers of devices using a single wireless power source .…

This paper studies the distributed linearly separable computation problem . In this problem, a master asks $N$ distributed workers to compute a function of $K$ datasets . We assign some datasetsto each worker, which then computes the corresponding messages and returns somefunction of these messages .…

We consider the setting of distributed storage system where a single file issubdivided into smaller fragments of same size . We model the fragment download time as an exponential random variable independent and identically distributed for all fragmentsacross all servers .…

In nature and engineering world, acquired signals are usually affected by multiple complicated factors and appear as multicomponent nonstationary modes . We use a kernel function-based complex quadratic phase function to represent a multicomponents signal in the three-dimensional space of time, frequency and chirp rate .…

Simflowny is an open platform which automatically generates efficient parallel code of scientific dynamical models for different simulationframeworks . Here we present major upgrades on this software to supports a quite generic family of partial differential equations . Theseequations can be discretized using: (i) standard finite-difference for systemswith derivatives up to any order, (ii) High-Resolution-Shock-Capturing methodsto deal with shocks and discontinuities of balance law equations .…

Meshless solution to differential equations using radial basis functions(RBF) is an alternative to grid based methods commonly used . Gaussian, Multiquadrics are some of the most popularRBFs used for the solutions of fluid flow and heat transfer problems . Polyharmonic splines (PHS) with appendedpolynomials have been shown to solve the above issues and give rapidconvergence of discretization errors with the degree of appended polynomials .…

A numerical method using discontinuous polynomial approximations isformulated for solving a phase-field model of two immiscible fluids with asoluble surfactant . The scheme recovers the Langmuir adsorption isotherms atequilibrium . The numerical method is used to simulate fluidflows in the pore space of Berea sandstone obtained by micro-CT imaging .…

We propose and rigorously analyze a mixedfinite element method for the approximation of the Hamilton–Jacobi–Bellman equation with coefficients satisfying the Cordes condition . The second part of the paper focuses on the numerical homogenization of such equations . Numerical experiments demonstrate the approximationscheme for the effective Hamiltonian and the numerical solution of the problem of the homogenized problem .…

We introduce a new general framework for the approximation of evolutionequations at low regularity . We develop a new class of schemes for a wide range of equations under lower regularity assumptions than classical methods require . We apply our general theory to the time discretization of concrete PDEs, such as the nonlinear heat equation, and the complex Ginzburg-Landau equation, the half wave and Klein–Gordon equations, set in $\Omega \subset \mathbb{R}^d$, $d \leq 3$ with suitable boundary conditions .…

We propose a unified framework for multi-view subspace learning to learn orthogonal projections for all views . The framework integrates thecorrelations within multiple views, supervised discriminant capacity, and distance preservation in a concise and compact way . An efficientnumerical method based on successive approximations via eigenvectors is presented to solve the associated optimization problem .…

Inverse problems are more challenging when only partial data are available ingeneral . In this paper, we propose a two-step approach combining the extended sampling method and the ensemble Kalman filter to reconstruct an elastic rigidobstacle using partial data . In the first step, the approximate location of the unknown obstacle is obtained by the extended sample method .…

The band structure problem is first formulated as the eigenvalueproblem of an operator function . Finite elements are then used for retraction . The spectral indicator method is employed to computethe eigenvalues . Numerical examples in both the TE and TM cases are presented to show the effectiveness of the method .…

We develop a one step matrix method in order to obtain approximate solutions of first order systems and non-linear ordinary differential equations . We find a sequence of such solutions that converge to the exact solution . We study the precision, in terms of the precision of the method by applying it to different well known examples .…

This chapter deals with multi-material flow problems by a kind of effectivenumerical methods, based on a series of reduced forms of the Baer-Nunziato (BN)model . We propose a new non-oscillatory Godunov-type scheme for computing multi-fluid flows in the Eulerian framework .…

When solving the Poisson equation by the finite element method, we use onedegree of freedom for interpolation by the given Laplacian – the right handside function in the partial differential equation . The finite element solutionis the Galerkin projection in a smaller vector space .…

We study a family of $H^m$-conforming piecewise polynomials based on artificial neural network, named as the finite neuron method (FNM) fornumerical solution of $2m-th order partial differential equations in $\mathbb{R}^d$ for any $m,d \geq 1$ We then provide convergence analysis for this method .…

This paper introduces a multi-level (m-lev) mechanism into EvolutionStrategies (ESs) in order to address a class of global optimization problemsthat could benefit from fine discretization of their decision variables . Problems arise in engineering and scientific applications, which possess amulti-resolution control nature, and thus may be formulated either by means of low-resolution variants (providing coarser approximations with presumably lower accuracy for the general problem) or by high-resolution controls .…

Hopfield neural networks are a possible basis for modelling associative memory in living organisms . We discuss the role of biases (the external inputs) in the learning process in Hopfield networks . We apply Newtons method for learning memories, and experimentally compare the performances of variouslearning rules .…

Detecting covert channels among legitimate traffic represents a severe challenge due to the high heterogeneity of networks . The proposed solution has been evaluated over a 15-day-long experimental session with the injection of traffic that covers the most relevant exfiltration and tunneling attacks .…

The goal is to find a program that satisfies user-provided constraints while maximizing the program’s score with respect to a neural model . At the core of our method is a top-down recurrent neural model thatplaces distributions over abstract syntax trees conditioned on the NL input .…

Micro-core architectures combine many low memory, low power computing corestogether in a single package . These are attractive for use as accelerators but due to limited on-chip memory and multiple levels of memory hierarchy, the way programmers offload kernels needs to be carefully considered .…

An extended Kalman filter is used to estimate the current location of target from its visual information . It then predicts its future position by using the observation sequence . The proposed algorithm is developed in ROS-Gazeboenvironment and is verified using hardware implementation .…

This work details the problem of aerial target capture using multiple UAVs . The problem is motivated from the challenge 1 of Mohammed Bin ZayedInternational Robotic Challenge 2020 . Multi-UAV collaboration improves theefficiency of the system and increases the chance of capturing the ball in short span of time .…

State-of-the-artday-time depth estimation methods fail miserably when tested with night-time images due to a large domain shift between them . We propose to solve this problem byposing it as a domain adaptation problem where a network trained with day-timeimages is adapted to work for night time images .…

Reliability in Neural Networks is crucial in safety-critical applications like healthcare . We propose an uncertainty-aware boosting technique for multi-modal ensembling topredict Alzheimer’s Dementia Severity . This work aims to encourage fair and aware models . The source code is available at https://://://github.com/wazeerzulfikar/alzheimers-dementia-sustaining-a-solution-alzheimer-scientist-predicting-suspectia…

Satellites and other complex systems now become more and more softwareddependent . Even nanosatellites have complexity that can be compared to scientific instruments launched to Mars . There is not a rigorous and robust way to design software for CubeSats or small satellites yet .…

Learning-based automatic program repair techniques are showing promise to provide quality fix suggestions for detected bugs in the source code . With the increasing popularity of code review, dependency on bug localizers can be reduced . The code review-based bug localization is more trustworthy since reviewers’ expertise and experience are reflected inthese suggestions .…

This paper proposes a procedure to execute external source codes from a LaTeX document and include the calculation outputs in the resulting Portable Document Format (pdf) file automatically . It integrates programming tools into the LaTeX tool to facilitate the production of reproducible research .…

Leadership and followership are essential parts of collective decision and organization in social animals, including humans . In nature, relationships of leaders and followers are dynamic and vary with context or temporal factors . In this paper, we focus on mining and modeling frequentpatterns of leading and following .…

Conferences are an essential tool for scientific communication . Over 50% of original research results are published in conference proceedings . The comparison of this bibliometric-driven ranking with the expert-driven CORE ranking in Computer Science showed a 62% overlap, as well as a significant average rank correlation of the category distribution .38 conference proceedings in Engineering (45% of the list) and 23 in ComputerScience (32%) have an SJR level corresponding to the first quartile journals in these areas .…