## Introducing Inter Relatedness between Wikipedia Articles in Explicit Semantic Analysis

Explicit Semantic Analysis (ESA) is a technique used to represent a piece of text as a vector in the space of concepts, such as Articles found in Wikipedia . We propose a methodology to incorporate knowledge of Inter-relatedness betweenWikipedia Articles to the vectors obtained from ESA using a technique calledRetrofitting to improve the performance of subsequent tasks that use ESA to form vector embeddings .…

## Fair Multi Stakeholder News Recommender System with Hypergraph ranking

Recommender systems are typically designed to fulfill end user needs . But in some domains users are not the only stakeholders in the system . In a news aggregator website users, authors, magazines as well asthe platform itself are potential stakeholders .…

## Fast Class wise Updating for Online Hashing

Online image hashing has received increasing research attention recently, which processes large-scale data in a streaming fashion to update the hashfunctions on-the-fly . A novel supervised online hashing scheme, termed Fast Class-wiseUpdating for Online Hashing, is proposed to address the above two challenges by introducing a novel and efficient inner product operation .…

## Diversifying Relevant Phrases

Diverse keyword suggestions for a given landing page or matching queries todiverse documents is an active research area in online advertising . Modern search engines provide advertisers with products like Dynamic Search Ads andSmart Campaigns where they extract meaningful keywords/phrases from theadvertiser’s product inventory .…

## Minimum ell_1 norm interpolation via basis pursuit is robust to errors

This article studies basis pursuit, i.e. minimum $\ell_1$-norm interpolation,in sparse linear regression with additive errors . It is assumed that the number of Gaussian features growssuperlinear in number of samples . The Euclidean error of recovering the true regressor is of the order of the average noise level .…

## Debiasing Evaluations That are Biased by Evaluations

It is common to evaluate a set of items by soliciting people to rate them . In these applications, students often give a higher rating to a course if they receive higher grades in a course . In this work, we call these external factors the “outcome” experienced by people .…

## Further strengthening of upper bounds for perfect k Hashing

An important asymptotic upper bound for general $k$ was derived by Fredmanand Koml\’os in the ’80s . In this paper, we further improve the bounds for $5\leq k \leq 8$ Themethod we use, which depends on the reduction of an optimization problem to afinite number of cases, shows that further results might be obtained by refined arguments at the expense of higher complexity .…

## Hardware Implementation of Iterative Projection Aggregation Decoding of Reed Muller Codes

In this work, we present a simplification and a corresponding hardwarearchitecture for hard-decision recursive projection-aggregation (RPA) decoding of Reed-Muller codes . We present FPGAim implementation results for an RM(6,3) code on a Xilinx Virtex-7 FPGA showing that our proposed architecture achieves a throughput of 171 Mbps at a frequency of 80 MHz .…

## Trade offs in Decentralized Multi Antenna Architectures The WAX Decomposition

As the number of antennas increases the computational complexity required to process data using centralizedarchitectitectures become prohibitively high . Decentralized architectures can reduce the requirements by pre-processing the data before it arrives at a central processing unit (CPU) However, performing decentralized processing introducesalso cost in complexity/interconnection bandwidth at the antenna end which isin general being ignored .…

## SuperCoder Program Learning Under Noisy Conditions From Superposition of States

A new method of program learning in a Domain Specific Language is based on gradient descent with no direct search . The method surpasses the state-of-the-art forsynthesising long programs and is able to learn programs under noise . The first component of our method is anattention-based recurrent neural network, which provides an appropriateinitialization point for the gradient descent that optimizes the probabilisticrepresentation of the DSL variables .…

## Combining nondeterminism probability and termination equational and metric reasoning

We study monads resulting from the combination of nondeterministic andprobabilistic behaviour with the possibility of termination . Our main contributions are presentation results for themonads providing equational reasoning tools for establishing equivalences and distances of programs .…

## Difference restriction algebras of partial functions with operators discrete duality and completion

We exhibit an adjunction between abstract algebras of partialfunctions and a category of set quotients . We show that the monad induced by our adjunctionyields the compatible completion of any atomic representable algebra . We then extend these adjunction,duality, and completion results to representable algebraas equipped witharbitrary additional completely additive and compatibility preservingoperators .…

## A new operational representation of dependencies in Event Structures

The execution of an event in a complex and distributed system can be represented in many ways . The aim of this paper is to propose whatcan be the appropriate kind of Petri net corresponding to Context-Dependent Event structures . Dependencies are often operationallyrepresented, in Petri nets, by tokens produced by activities and consumed by others .…

## Cost Efficient Repository Management for Cloud Based On Demand Video Streaming

Video transcoding is the process of converting a video to the formats supported by the viewer’s device . Video streaming providers generally need to prepare several formats of the same video (termed pre-transcoding) and stream the appropriateformat to the viewer .…

## Ultra low bitrate video conferencing using deep image animation

In this work we propose a novel deep learning approach for ultra-low bitrate video compression for video conferencing applications . We adopt a model-based approach that employs deep neural networks to encode motion information as keypoint displacement and construct the video signal at the decoder side .…

## Dynamic Image for 3D MRI Image Alzheimer s Disease Classification

We propose to apply a 2D CNN architecture to 3D MRI image Alzheimer’s disease classification . We show our proposed CNN model achieves $9.5\%$ better Alzheimer’sdisease classification accuracy than the baseline 3D models . We also show thatour method allows for efficient training, requiring only 20% of the trainingtime .…

## A robust and scalable unfitted adaptive finite element framework for nonlinear solid mechanics

We extend the unfitted $h$-adaptive Finite Element Method to handle nonlinear problems in solid mechanics . The method is implementedconsidering both irreducible and mixed (u/p) formulations . In the numerical experiments, both formulations are used to model the inelastic behavior of compressible and incompressible materials .…

## Combined Sieve Algorithm for Prime Gaps

A new Combined Sieve algorithm is presented with cost proportional to thenumber of enumerated factors over a series of intervals . This algorithmachieves a significant speedup, over a traditional sieve, when handling many [10^4, 10^7] intervals concurrently . In real world tests, this new algorithm regularly runs 10,000x faster .…

## 6 7ms on Mobile with over 78 ImageNet Accuracy Unified Network Pruning and Architecture Search for Beyond Real Time Mobile Acceleration

With the increasing demand to efficiently deploy DNNs on mobile edge devices,it becomes much more important to reduce unnecessary computation and increasethe execution speed . Prior methods towards this goal, including modelcompression and network architecture search (NAS), are largely performedindependently .…

## Solvable Model for Inheriting the Regularization through Knowledge Distillation

Knowledge distillation where a smallerneural network is trained using the outputs of a larger neural network is aparticularly interesting case of transfer learning . We areable to show that, through KD, the regularization properties of the largerteacher model can be inherited by the smaller student .…

## A Comparative Evaluation of Population based Optimization Algorithms for Workflow Scheduling in Cloud Fog Environments

This work presents a comparative evaluation of four population-basedoptimization algorithms for workflow scheduling in cloud-fog environments . These algorithms are as follows: Particle Swarm Optimization (PSO), GeneticAlgorithm (GA), Differential Evolution (DE) and GA-PSO . Future work will include expansion of the workflows used by increasing the number of tasks as well as adding some more workflows .…

## Opportunistic Routing Metrics A Timely One Stop Tutorial Survey

High-speed, low latency, and heterogeneity features of 5G, as the commondenominator of many emerging and classic wireless applications, have put wireless technology back in the spotlight . The task of choosing or designing anappropriate OR metric is a critical one .…

## Edge assisted Democratized Learning Towards Federated Analytics

Federated Analytics (FA) reuses the Federated Learning (FL) infrastructure to evaluate the population-level summary of model performances . The current realization of FL adopts single server-multiple clientarchitecture with limited scope for FA . We need to design a robust learning mechanism than the FL that unleashes a viable infrastructure for FA and trains learning models with better generalization capability .…

## Design and Implementation of Fair Congestion Control for Data Centers Networks

In data centers, the nature of the composite bursty traffic along with the small bandwidth-delay product and switch buffers lead to several congestion problems that are not handled well by traditional congestion control mechanism . We present a theoretical analysis to show the stability and effectiveness of the scheme .…

## Designing Voice Controllable APIs

The main purpose of a voice command system is to process a sentence innatural language and perform the corresponding action . This mapping is usually performed after the API is already implemented, possibly by other programmers . In this paper, we describe how the API developercan use patterns to map sentences to API calls by utilizing the similarities between names and types in the sentences and the API .…

## TracerX Dynamic Symbolic Execution with Interpolation

Dynamic Symbolic Execution (DSE) is an important method for the testing of programs . We present anew interpolation algorithm and implement it on top of the KLEE system . The main objective is to address the path explosion problem in pursuit of codepenetration: to prove that a target program point is either reachable orunreachable .…

## Deep Ad hoc Beamforming Based on Speaker Extraction for Target Dependent Speech Separation

Ad-hoc microphone array may cover such a large area that multiple speakers may locate far apart and talk independently . We propose a supervised channel selection framework based on speakerextraction, where the estimated utterance-level SNRs of the target speech are used as the basis for the channel selection .…

## Improving accuracy of rare words for RNN Transducer through unigram shallow fusion

Unigram shallow fusion (USF) can improve performance on rare words by 3.7% WER relative without degradation on general test set . USF works by fixing errors in probability estimates of words due to Viterbi search used during decoding with subword-based RNN-T .…

## HPM Frame A Decision Framework for Executing Software on Heterogeneous Platforms

Heterogeneous computing is one of the most important computational solutionsto meet rapidly increasing demands on system performance . The refactoring of systems forexecution on such platforms is highly desired but also difficult to perform . We have identified a current need for asystematic approach that supports engineers in the refactororing process — from CPU-centric applications to software that is executed on heterogeneousplatforms .…

## Software Security Patch Management A Systematic Literature Review of Challenges Approaches Tools and Practices

Software security patch management purports to support the process of patching known software security vulnerabilities . Only 20.8% of thereported solutions have been rigorously evaluated in industrial settings . Two-thirds of the common challenges havenot been directly addressed in the solutions .…

## Design Level Metrics to Measure the Complexity Across Versions of AO Software

Software metric plays a vital role in quantitative assessment of any specific software development methodology and its impact on the maintenance of software . It can also indicate the degree of interdependence among the components by providing valuable feedback about quality attributes such asmaintainability, modifiability and understandability .…

## Evaluation of Reusability in Aspect Oriented Software using Inheritance Metrics

Aspect-Oriented Software Development (AOSD) is a promising methodology forefficiently capturing the cross-cutting functionalities (concerns) as independent units called aspects . Inheritance of classes and aspects play avital role in defining the units of encapsulation . It is essential toquantitatively capture the impact of inheritance in AOSD using design levelmetrics and to infer on the higher level quality attribute, reusability .…

## Vertex Sparsification for Edge Connectivity in Polynomial Time

An important open question is whether $(1+\epsilon)$-approximate cut-preserving vertices with size close to the number of terminals exist . We show thatconnectivity-$c$ mimicking networks with $\widetilde{O}(kc^3)$ edges exist and can be constructed in polynomial time in $n$ and $c$ improving over the results of Chalermsook et al.…

## Parameterized algorithms for locating dominating sets

A locating-dominating set of a graph is a dominating set of $G$ whereeach vertices not in $D$ has a unique neighborhood . This problem is known to be $mathsf{NP-hard}$ even on restrictedgraphs, such as interval graphs, split graphs, and planar bipartitesubcubic graphs .…