Abstract

In the states of Andhra Pradesh and Telangana in India, the number of engineering institutions grew from under 10 to over 500 from 1978 to 2008. This resulted in a severe shortage of good quality teachers and therefore poor quality of graduating students. In India, the employability of engineering graduates is 25%. Considering the scale and wickedness of the problem, the lead author and her team worked on designing innovative and radical solutions to this problem during the last eight years. We restrict our research focus to those colleges who responded to our call for partnership in Andhra Pradesh and Telangana and to the center, EnhanceEdu, we created at IIIT (International Institute of Information Technology) at Hyderabad. We present the EnhanceEdu design story using a new pragmatic approach called Design Story Research (DeStoRe) which has its basis in Design Science Research. An overarching design story frames past, present and future work of an entity, seeing possibilities for radical innovation

Abstract

Unsupervised representation learning for tweets is an important research field which helps in solving several business applications such as sentiment analysis, hashtag prediction, paraphrase detection and microblog ranking. A good tweet representation learning model must handle the idiosyncratic nature of tweets which poses several challenges such as short length, informal words, unusual grammar and misspellings. However, there is a lack of prior work which surveys the representation learning models with a focus on tweets. In this work, we organize the models based on its objective function which aids the understanding of the literature. We also provide interesting future directions, which we believe are fruitful in advancing this field by building high-quality tweet representation learning models.

Abstract

Web is indeed prolific with useful content. However, some segments of the user population – for example, students of less rigorous colleges from rural India – may find some of the relevant material on the web complex, foreign and incomprehensible. So, while good content exists, not all are able to benefit from it. The intent of our research effort is to be able to make existing web content be more accessible to wider range of end-users who may have different sociocultural and cognitive abilities. We propose to overcome this gap at a Presentation Semantics level by developing a framework for renarration through Semantic Style Sheets. Using this framework, we facilitate the renarration of existing educational content to be used by thus far overlooked community of users.

Abstract

Owing to the tremendous increase in the volume and variety of user generated content, train–once–apply–forever models are insufficient for supervised learning tasks. Œe need is to develop algorithms that can adapt across domains by leveraging labeled data from source domain(s) and efficiently perform the task in the unlabeled target domain. Towards this, we present a novel two-stage neural network learning algorithm for domain adaptation which learns a multi-part hidden layer where individual parts contribute dif- ferently to the tasks in source and target domains. Œe multiple parts of the representation (i.e. hidden layer) are learned while being cognizant of what characteristics to transfer across domains and what to preserve within domains for enhanced performance. Œe first stage embroils around learning a two-part representation i.e. source specific and common representations in a manner such that the former do not detract the ability of the later to represent the target domain. In the second stage, the generalized common representation is further iteratively extended with discriminating target specific characteristics to adapt to the target domain. We em- pirically demonstrate that the learned representations, in different arrangements, outperform existing domain adaptation algorithms in the source classification as well as the cross-domain classification tasks o

Abstract

HTTPS error warnings are supposed to alert browser users to network attacks. Unfortunately, a wide range of non-attack circumstances trigger hundreds of millions of spurious browser warnings per month. Spurious warnings frustrate users, hinder the widespread adoption of HTTPS, and undermine trust in browser warnings. We investigate the root causes of HTTPS error warnings in the field, with the goal of resolving benign errors. We study a sample of over 300 million errors that Google Chrome users encountered in the course of normal browsing. After manually reviewing more than 2,000 error reports, we developed automated rules to classify the top causes of HTTPS error warnings. We are able to automatically diagnose the root causes of two-thirds of error reports. To our surprise, we find that more than half of errors are caused by client-side or network issues instead of server misconfigurations. Based on these findings, we implemented more actionable warnings and other browser changes to address clientside error causes. We further propose solutions for other classes of root causes.

Abstract

A dominating set of a small size is useful in several settings including wireless networks, document summarization, secure system design, and the like. In this paper, we start by studying three distributed algorithms that produce a small sized dominating sets in a few rounds. We interpret these algorithms in the natural shared memory setting and experiment with these algorithms on a multi-core CPU. Based on the observations from these experimental results, we propose variations to the three algorithms and also show how the proposed variations offer interesting trade-offs with respect to the size of the dominating set produced and the time taken.

Abstract

Advertisements (ads) often include strongly emotional content to leave a lasting impression on the viewer. This work (i) compiles an affective ad dataset capable of evoking coherent emotions across users, as determined from the affective opinions of five experts and 14 annotators; (ii) explores the efficacy of convolutional neural network (CNN) features for encoding emotions, and observes that CNN features outperform low-level audio-visual emotion descriptors[9] upon extensive experimentation; and (iii) demonstrates how enhanced affect prediction facilitates computational advertising, and leads to better viewing experience while watching an online video stream embedded with ads based on a study involving 17 users. We model ad emotions based on subjective human opinions as well as objective multimodal features, and show how effectively modeling ad emotions can positively impact a real-life application.

Abstract

Despite the fact that advertisements (ads) often include strongly emotional content, very little work has been devoted to affect recognition (AR) from ads. This work explicitly compares content-centric and user-centric ad AR methodologies, and evaluates the impact of enhanced AR on computational advertising via a user study. Specifically, we (1) compile an affective ad dataset capable of evoking coherent emotions across users; (2) explore the efficacy of content-centric convolutional neural network (CNN) features for encoding emotions, and show that CNN features outperform low-level emotion descriptors; (3) examine user-centered ad AR by analyzing Electroencephalogram (EEG) responses acquired from eleven viewers, and find that EEG signals encode emotional information better than content descriptors; (4) investigate the relationship between objective AR and subjective viewer experience while watching an ad-embedded online video stream based on a study involving 12 users. To our knowledge, this is the first work to (a) expressly compare user vs content-centered AR for ads, and (b) study the relationship between modeling of ad emotions and its impact on a real-life advertising application.

Abstract

Quality control of responses enhances sustainability and adoption of crowdsourcing. Expert and peer reviews, majority voting, machine learning, game theory, etc. are some of the practices for quality control in crowdsourcing. However, quality of crowdsourced responses is still a concern. We propose a quality control approach drawing inspiration from Requirements Engineering quality attributes - Completeness, Consistency and Correctness (3Cs). The 3Cs of a response are assessed and displayed as a CSQuaRE score based on coverage with reference to knowledge base (ontology), cohesiveness and contributor credibility in the domain. The knowledge base would evolve with continued extraction of instances of information and thus responses would be re-calibrated for relevance. The suggested approach would be demonstrated for Information Security related Question and Answers on a crowdsourcing platform. The evaluation of the approach would be based on comparison with existing quality control techniques and feedback from security experts.

Abstract

Recent work on Acoustic Unit Discovery (AUD) has led to the development of a non-parametric Bayesian phone-loop model where the prior over the probability of the phone-like units is assumed to be sampled from a Dirichlet Process (DP). In this work, we propose to improve this model by incorporating a Hierarchical Pitman-Yor based bigram Language Model on top of the units' transitions. This new model makes use of the phonotactic context information but assumes a fixed number of units. To remedy this limitation we first train a DP phone-loop model to infer the number of units, then, the bigram phone-loop is initialized from the DP phone-loop and trained until convergence of its parameters. Results show an absolute improvement of 1–2%on the Normalized Mutual Information (NMI) metric. Furthermore, we show that, combined with Multilingual Bottleneck (MBN) features the model yields a same or higher NMI as an English phone recogniser trained on TIMIT.