ICodeNet consists of an ImageNet trainedVGG encoder followed by a shallow neural network . The shallow network is based on either CNN or LSTM . Different variations of models are evaluated on a sourcecode author classification dataset . We have also compared our image-basedhierarchical neural network model with simple image-freezing CNN architecture and text-based CNN and L STM models to highlight its novelty and efficiency.…
Near field Image Transmission and EVM Measurements in Rich Scattering Environment
In this work, we present near-field image transmission and error vectormagnitude measurement in a rich scattering environment in a metal enclosure . Wecheck the effect of loading metal enclosure on the performance of SDR basednear-field communication link . Thenear-field performance is measured by transmitting wideband OFDM-modulatedpackets containing image information .…
PV RCNN Point Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection
3D object detection is receiving increasing attention from both industry andacademia thanks to its wide applications in various fields . In this paper, we propose the Point-Voxel Region based Convolution Neural Networks (PV-RCNNs) foraccurate 3D detection from point clouds . We propose a more advanced framework for more efficient and accurate 3D discovery .…
Admix Enhancing the Transferability of Adversarial Attacks
Admix AttackMethod (AAM) is a new input transformation based attack based attack . It considers both the original image and an image randomly picked from other categories . AAM calculates the gradient on the admixed image in order to craft adversaries with higher transferablility .…
A Simple yet Brisk and Efficient Active Learning Platform for Text Classification
In this work, we propose the use of a fully managed machine learning service, which utilizes active learning to directly build models from unstructured data . Our approach leverages state-of-the-art text representation like OpenAI’s GPT2 and a fast implementation of the active learning workflow that relies on a simple construction of incremental learning using linear models .…
Classification of Fracture and Normal Shoulder Bone X Ray Images Using Ensemble and Transfer Learning With Deep Learning Models Based on Convolutional Neural Networks
Shoulder boneXray images were classified and compared via deep learning models based onconvolutional neural network (CNN) using transfer learning and ensemblelearning . The study was conducted using all shoulder images in musculoskeletal radiographs (MURA) dataset, one of the largest publicradiographic image datasets .…
A Supervised Learning Approach for Robust Health Monitoring using Face Videos
Monitoring of cardiovascular activity is highly desired and can enable novel applications in diagnosing potential cardiovascular diseases and maintaining an individual’s well-being . Currently, such vital signs are measured using contact devices such as an electrocardiogram (ECG), chest straps, and pulse oximeters that require the patient or the health provider to manually implement .…
Graph Neural Networks to Predict Customer Satisfaction Following Interactions with a Corporate Call Center
The system takes as an input speech-to-text transcriptions of calls and predicts call satisfactionreported by customers on post-call surveys (scale from 1 to 10) Because of its subjective nature, predicting survey scores is not a trivial task and presents several modeling challenges .…
Characterizing Student Engagement Moods for Dropout Prediction in Question Pool Websites
Problem-Based Learning (PBL) is a popular approach to instruction that supports students to get hands-on training by solving problems . But empirical findings suggest that 40% to 80% of students registered in QPs drop out in less than two months .…
A Multiscale Environment for Learning by Diffusion
The Multiscale Environment for Learning by Diffusion (MELD) datamodel is a family of clusterings parameterized by nonlinear diffusion on the dataset . We show that the MELD data model precisely captures latent multiscale structure in data and facilitates its analysis .…
Co Seg An Image Segmentation Framework Against Label Corruption
Co-Seg is a novel deeplearning framework to train segmentation networks on noisy datasets . It trains two networks simultaneously to sift through all samples and obtain reliable labels . Then an efficient yet easily-implemented labelcorrection strategy is applied to enrich the reliable subset .…
On the maximum number of non attacking rooks on a high dimensional simplicial chessboard
Martin and Wagner (Graphs Combin.(2015) 31:1589–1611) asked about the independence number of ${\rm\mathcal{SR}}(m,n)$ that is the maximum number of non attacking rooks which can be placed on a $m-1-dimensional simplicial chessboard of side length $n+1$ Martin and W. Wagner solve this problem and show that $alpha(1-o(1) =\rm) and $alpha (1) is $1) .…
A Trust Based Approach for Volunteer Based Distributed Computing in the Context of Biological Simulation
A trust-based approach uses trusted communities that are a subset of all clients where they trust each other . Using such TCs, the system becomes more organic and responds better to malicious or malfunctioning clients . This approach is more organic than previous attempts to simulate complex biological processes, such as the creation of a computer that simulates biological processes such as those of nature .…
New Type I Binary 72 36 12 Self Dual Codes from Composite Matrices and R1 Lifts
In this work, we define three composite matrices derived from group rings . We find 30 new Type I binary self-dual codes with parameters [36, 18, 6 or 8]. Many of these codes turn out to have weight enumerators with parameters that were not known in the literature before .…
On the Purity of Resolutions of Stanley Reisner Rings Associated to Reed Muller Codes
Following Johnsen and Verdure (2013), we can associate to any linear code $C$an abstract simplicial complex and in turn a Stanley-Reisner ring $R_C$ . The question of purity of the minimalfree resolution was considered by Ghorpade and Singh (2020) They showed that the resolution is pure insome cases and it is not pure in many other cases .…
Towards Imperceptible Query limited Adversarial Attacks with Perceptual Feature Fidelity Loss
Recently, there has been a large amount of work towards fooling deep-learning-based classifiers . However, researchers usually use Lp-norm minimization as a proxy for imperceptibility, which oversimplifies the diversity and richness of real-world images and human visual perception . The metric is particularly useful inthe challenging black-box attack with limited queries, where theimperceptibility is hard to achieve due to the non-trivial perturbation power.…
Shapley Scarf Housing Markets Respecting Improvement Integer Programming and Kidney Exchange
In a housing market of Shapley and Scarf, each agent is endowed with oneindivisible object and has preferences over all objects . An allocation of the objects is in the (strong) core if there exists no (weakly) blocking coalition . If an agent’sobject becomes more attractive for some other agents, then the agent’sallotment in the unique strong core allocation weakly improves .…
BNLP Natural language processing toolkit for Bengali language
BNLP is an open source language processing toolkit for Bengali language . It provides pre-trained model with high accuracy to do modelbased tokenization, embedding, POS tagging, NER tagging task for Bengalilanguage . BnLP is usingwidely in the Bengali research communities with 16K downloads, 119 stars and 31forks .…
Computation of multi degree Tchebycheffian B splines
Multi-degree Tchebycheffian splines are splines with pieces drawn fromextended (complete) Tche bycheff spaces . These are a natural extension of multi-degree polynomial splines . Wepresent a practical framework to compute MDTB-splines, and provide anobject-oriented implementation in Matlab . The implementation supports the construction, differentiation, and visualization of MDTB splines whose piecesbelong to TcheByCheff spaces that are null-spaces of constant-coefficientlinear differential operators .…
A Quantum Interpretation of Bunched Logic for Quantum Separation Logic
We propose a model of the substructural logic of Bunched Implications (BI) that is suitable for reasoning about quantum states . We develop a program logic where pre- and post-conditions are BI formulas describing quantumstates . We exercise logic for proving the security of quantum one-time pad and secret sharing .…
The Zero Cubes Free and Cubes Unique Multidimensional Constraints
This paper studies two families of constraints for two-dimensional andmultidimensional arrays . The first family requires that a multidimensionalarray will not contain a cube of zeros of some fixed size . The second family imposes that there will not be two identical cubes of a given size .…
Exponential Savings in Agnostic Active Learning through Abstention
We show that in pool-based active classification without assumptions on the underlying distribution, if the learner is given the power to abstain from somepredictions by paying the price marginally smaller than the average loss $1/2$ of a random guess, exponential savings in the number of label requests are possible whenever they are possible in the corresponding realizable problem .…
Random walks and forbidden minors III poly d ε time partition oracles for minor free graph classes
An important tool for sublinear algorithmsand property testing for such classes is the partition oracle . The partitionoracle of Hassidim et al. runs in time d^poly(d/\e) per query . We build on a recent spectral graph theoretical toolkit forminor-closed graph families, introduced by the authors to develop efficientproperty testers .…
Open Source Concealed EEG Data Collection for Brain Computer Interfaces Real World Neural Observation Through OpenBCI Amplifiers with Around the Ear cEEGrid Electrodes
An integrated system design combines electronics components with newly designed 3D-printed partsto form an easily replicable, versatile, single-unit EEG recording system for prolonged use and easy application development . Support for alink between temporal Alpha power and flow is found, which indicates anefficient engagement of verbal-analytic reasoning with intensified flow experiences, and specifically intensified task absorption .…
Beyond the Command Feminist STS Research and Critical Issues for the Design of Social Machines
Machines, from artificially intelligent digital assistants to embodied robots, are becoming more pervasive in everyday life . This paper contributes the “social machine” as a model for technology designers who seek to recognize the importance, diversity and complexity of the social in their work .…
Group Matrix Ring Codes and Constructions of Self Dual Codes
In this work, we study codes generated by elements that come from groupmatrix rings . We present a matrix construction which we use to generate codesin two different ambient spaces: the matrix ring $M_k(R)$ and the ring $R,$where $R$ is the commutative Frobenius ring .…
Actor Critic Learning Based QoS Aware Scheduler for Reconfigurable Wireless Networks
The flexibility offered by reconfigurable wireless networks, provide new opportunities for various applications such as online AR/VR gaming, high-quality video streaming and autonomous vehicles, that desire high-bandwidth, reliable and low-latency communications . These applications come with very stringent Quality of Service (QoS) requirements and increase the burden over mobile networks .…
MRIReco jl An MRI Reconstruction Framework written in Julia
The aim of this work is to develop a high-performance, flexible andeasy-to-use MRI reconstruction framework using the scientific programminglanguage Julia . Julia is a modern, general purpose programming language with strongfeatures in the area of signal / image processing and numerical computing .…
EdgeWorkflowReal An Edge Computing based Workflow Execution Engine for Smart Systems
Current cloud-based smart systems suffer from weaknesses such as highresponse latency, limited network bandwidth and restricted computing power of smart end devices . EdgeWorkflowReal is an edgecomputing based workflow execution engine for smart systems . It is a challenging job for software developers who do not have the skills for thecreation of edge computing environments .…
A universal solution scheme for fractional and classical PDEs
We propose a unified meshless method to solve classical and fractional PDE problems with $–\Delta)^{\frac{frac{\alpha}{2} for $-alpha = 2$ and $-Alpha = 0$ . The method can achieve high accuracy with fewer number of unknowns . It bypasses numerical approximation to the hypersingular integral offractional Laplacian .…
Expressive Neural Voice Cloning
Voice cloning is the task of learning to synthesize the voice of an unseenspeaker from a few samples . We propose a controllable voice cloning method that allows fine-grained control over various style aspects of the synthesized speech for an unseen speaker .…
Regression or Classification New Methods to Evaluate No Reference Picture and Video Quality Models
Video and image quality assessment has long been projected as a regressionproblem, which requires predicting a continuous quality score given an inputstimulus . We propose two new methods – binary,and ordinal classification – as alternatives to evaluate and compare no-reference quality models at coarser levels .…
The Challenges of Assessing and Evaluating the Students at Distance
The COVID-19 pandemic has caused a strong effect on higher educationinstitutions with the closure of classroom teaching activities . This short essay aims toexplore the challenges posed to Portuguese higher education institutions and toanalyze the challenge posed to evaluation models .…
Latent Space Inpainting for Packet Loss Concealment in Collaborative Object Detection
Edge devices, such as cameras and mobile units, are increasingly capable of performing sophisticated computation in addition to their traditional roles insensing and communicating signals . The focus of this paper is on collaborativeobject detection, where deep features computed on the edge device from inputimages are transmitted to the cloud for further processing .…
BAMSim Simulator
Resource allocation is an essential design aspect in multi-protocol labelswitched and OpenFlow/SDN network infrastructures . The bandwidth allocationmodels (BAMs) are an alternative to allocate and share bandwidth among network users . BAMs have an extensive number of parameters that need to be defined andtuned to achieve an expected network performance .…
Directional Sparse Filtering using Weighted Lehmer Mean for Blind Separation of Unbalanced Speech Mixtures
In blind source separation of speech signals, the inherent imbalance in the spectrum poses a challenge for methods that rely on single-sourcedominance for the estimation of the mixing matrix . We propose an algorithmbased on the directional sparse filtering (DSF) framework that utilizes theLehmer mean with learnable weights to adaptively account for source imbalance .…
Low Power Audio Keyword Spotting using Tsetlin Machines
The emergence of Artificial Intelligence (AI) driven Keyword Spotting (KWS) has revolutionized human to machine interaction . Yet, the challenge of end-to-end energy efficiency, memory footprint and systemcomplexity of current Neural Network (NN) powered AI-KWS pipelines has remained present . This paper evaluates KWS utilizing a learning automata powered machine learning algorithm called the Tsetlin Machine (TM) Through significantreduction in parameter requirements and choosing logic over arithmetic based processing, the TM offers new opportunities for low-power KWS while maintaining learning efficacy .…
Computability Complexity Consistency and Controllability A Four C s Framework for cross disciplinary Ethical Algorithm Research
The ethical consequences of, constraints upon and regulation of algorithmsarguably represent the defining challenges of our age . In this work, we provide a framework to assistcross-disciplinary collaboration by presenting a Four C’s Framework covering key computational considerations researchers across such diverse fields shouldconsider when approaching these questions .…
A self supervised learning based 6 DOF grasp planning method for manipulator
To realize a robust robotic grasping system for unknown objects in anunstructured environment, large amounts of grasp data and 3D model data for the object are required . To reduce the time cost of data acquisition and labeling and increasethe rate of successful grasps, we developed a self-supervised learningmechanism to control grasp tasks performed by manipulators .…
On the Expressive Power of Homomorphism Counts
A classical result by Lov\’asz asserts that two graphs $G$ and $H$ areisomorphic if and only if they have the same left-homomorphism vector . Dell, Grohe, and Rattan showed that restrictions of the left-Homomorphism . to a class of graphs can capture several different relaxations of isomorphism, including co-spectrality(i.e.,…
NL CNN A Resources Constrained Deep Learning Model based on Nonlinear Convolution
A novel convolution neural network model, abbreviated NL-CNN is proposed . The code for its implementation and some trained models are made publicly available . The proposed network gives very good testing accuracy, given a low implementation complexity and model size .…
Distributed Control of Multi Robot Systems in the Presence of Deception and Denial of Service Attacks
This research proposes a distributed switching control to secure multi-robotsystems in the presence of cyberattacks . Two major types of cyberattack are considered: deception attack and denial of service (DoS) attack . The performance of the proposed approach isevaluated on the Robotarium multi-robot testbed .…
Certified evaluations of Hölder continuous functions at roots of polynomials
Various methods can obtain certified estimates for roots of polynomials . For analytic evaluationfunctions, Newton’s method naturally applies to yield certified estimates . These estimates no longer apply, however, for H\”older continuous functions, which are a generalization of Lipschitz continuous functions where continuousderivatives need not exist .…
Semi supervised Sound Event Detection using Random Augmentation and Consistency Regularization
Sound event detection is a core module for acoustic environmental analysis . Semi-supervised learning technique allows to largely scale up the dataset without increasing the annotation budget . Data augmentation is important for the success of recent deep learning systems, study finds .…
Adversarially learning disentangled speech representations for robust multi factor voice conversion
Conventionalspeech representation learning methods in VC only factorize speech as speakerand content . We propose adisentangled speech representation learning framework based on adversariallearning . Four speech representations characterizing content, timbre, rhythmand pitch are extracted . The adversarial network is used to minimize the correlations between the speech representations, by randomly masking and predicting one of the representations from the others.…
Settling the Sharp Reconstruction Thresholds of Random Graph Matching
This paper studies the problem of recovering the hidden vertices correspondence between two edge-correlated random graphs . We prove that there exists a sharp threshold, above which one can correctly match all but a vanishing fraction of vertices and below which matching any positive fraction is impossible .…
Context Responsive Labeling in Augmented Reality
Augmented Reality (AR) enables mobile usersto utilize text labels, in order to provide a composite view associated withadditional information in a real-world environment . Displaying all labels for points of interest on a mobile device will lead to unwanted overlaps between information and labels .…
Symmetry Aware Reservoir Computing
The symmetry-aware RC can obtain zero error using anexponentially reduced number of artificial neurons and training data, greatlyspeeding up the time-to-result . We anticipate that generalizations of our procedure will have widespread applicability in information processing withANNs . We use our method to the parity task, a challenging benchmark problem, which highlights the benefits of symmetry matching .…
Resource Availability in the Social Cloud An Economics Perspective
This paper focuses on social cloud formation, where agents are involved in acloseness-based conditional resource sharing and build their resource sharingnetwork themselves . The objectives of this paper are: (1) to investigate theimpact of agents’ decisions of link addition and deletion on their local and global resource availability, (2) to analyze spillover effects in terms of the impact of the link addition between a pair of agents on others’ utility, (3) to study the role of agents’ closeness in determining what type of spillovereffects these agents experience in the network, and (4) to model the choices of agents that suggest with whom they want to add links in the social cloud .…
Modeling how social network algorithms can influence opinion polarization
A model simulates the communication in an online social network, in which theposts are created from external information . The dynamics starts with a user that has contact with a random opinion, and, according to a givenprobability function, this individual can post this opinion .…