arXiv subjects

Tanmoy Chakraborty

Publications and source records attributed to Tanmoy Chakraborty.

At least 253 records · Page 14Linked to original sources

arXiv2019

Fiducia: A Personalized Food Recommender System for Zomato

This paper presents Fiducia, a food review system involving a pipeline which processes restaurant-related reviews obtained from Zomato (India's largest restaurant search and discovery service). Fiducia is specific to popular cafe food items and manages to identify relevant information pertaining to each item separately in the reviews. It uses a sentiment check on these pieces of text and accordingly suggests an appropriate restaurant for the particular item depending on user-item and item-item similarity. Experimental results show that the sentiment analyzer module of Fiducia achieves an accuracy of over 85% and our final recommender system achieves an RMSE of about 1.01 beating other baselines.

cs.IR↗

arXiv2019

Multi-task Learning for Target-dependent Sentiment Classification

Detecting and aggregating sentiments toward people, organizations, and events expressed in unstructured social media have become critical text mining operations. Early systems detected sentiments over whole passages, whereas more recently, target-specific sentiments have been of greater interest. In this paper, we present MTTDSC, a multi-task target-dependent sentiment classification system that is informed by feature representation learnt for the related auxiliary task of passage-level sentiment classification. The auxiliary task uses a gated recurrent unit (GRU) and pools GRU states, followed by an auxiliary fully-connected layer that outputs passage-level predictions. In the main task, these GRUs contribute auxiliary per-token representations over and above word embeddings. The main task has its own, separate GRUs. The auxiliary and main GRUs send their states to a different fully connected layer, trained for the main task. Extensive experiments using two auxiliary datasets and three benchmark datasets (of which one is new, introduced by us) for the main task demonstrate that MTTDSC outperforms state-of-the-art baselines. Using word-level sensitivity analysis, we present anecdotal evidence that prior systems can make incorrect target-specific predictions because they miss sentiments expressed by words independent of target.

cs.LG↗

arXiv2019

Heterogeneous Edge Embeddings for Friend Recommendation

We propose a friend recommendation system (an application of link prediction) using edge embeddings on social networks. Most real-world social networks are multi-graphs, where different kinds of relationships (e.g. chat, friendship) are possible between a pair of users. Existing network embedding techniques do not leverage signals from different edge types and thus perform inadequately on link prediction in such networks. We propose a method to mine network representation that effectively exploits heterogeneity in multi-graphs. We evaluate our model on a real-world, active social network where this system is deployed for friend recommendation for millions of users. Our method outperforms various state-of-the-art baselines on Hike's social network in terms of accuracy as well as user satisfaction.

cs.SI↗

arXiv2018

Investigation of Thermodynamic Properties of Cu(NH3)4SO4.H2O, a Heisenberg Spin Chain Compound

Detailed experimental investigations of thermal and magnetic properties are presented for Cu(NH3)4SO4.H2O, an ideal uniform Heisenberg spin half chain compound. A comparison of these properties with relevant spin models is also presented. The temperature dependent magnetic susceptibility and specific heat data has been compared with the exact solution for uniform Heisenberg chain model derived by means of Bethe ansatz technique. Field dependent isothermal magnetization curves are simulated by Quantum Monte Carlo technique and compared with the corresponding experimental ones. Specific heat as a function of magnetic field (up to 7T) and temperature (down to 2K) is reported. Subsequently, the data are compared with the corresponding theoretical curves for the infinite Heisenberg spin half chain model with J=6K. Moreover, internal energy and entropy are calculated by analyzing the experimental specific heat data. Magnetic field and temperature dependent behavior of entropy and internal energy are in good agreement with the theoretical predictions.

cond-mat.str-el↗

arXiv2018

GIRNet: Interleaved Multi-Task Recurrent State Sequence Models

In several natural language tasks, labeled sequences are available in separate domains (say, languages), but the goal is to label sequences with mixed domain (such as code-switched text). Or, we may have available models for labeling whole passages (say, with sentiments), which we would like to exploit toward better position-specific label inference (say, target-dependent sentiment annotation). A key characteristic shared across such tasks is that different positions in a primary instance can benefit from different `experts' trained from auxiliary data, but labeled primary instances are scarce, and labeling the best expert for each position entails unacceptable cognitive burden. We propose GITNet, a unified position-sensitive multi-task recurrent neural network (RNN) architecture for such applications. Auxiliary and primary tasks need not share training instances. Auxiliary RNNs are trained over auxiliary instances. A primary instance is also submitted to each auxiliary RNN, but their state sequences are gated and merged into a novel composite state sequence tailored to the primary inference task. Our approach is in sharp contrast to recent multi-task networks like the cross-stitch and sluice network, which do not control state transfer at such fine granularity. We demonstrate the superiority of GIRNet using three applications: sentiment classification of code-switched passages, part-of-speech tagging of code-switched text, and target position-sensitive annotation of sentiment in monolingual passages. In all cases, we establish new state-of-the-art performance beyond recent competitive baselines.

cs.LG↗

arXiv2018

CQASUMM: Building References for Community Question Answering Summarization Corpora

Community Question Answering forums such as Quora, Stackoverflow are rich knowledge resources, often catering to information on topics overlooked by major search engines. Answers submitted to these forums are often elaborated, contain spam, are marred by slurs and business promotions. It is difficult for a reader to go through numerous such answers to gauge community opinion. As a result summarization becomes a prioritized task for CQA forums. While a number of efforts have been made to summarize factoid CQA, little work exists in summarizing non-factoid CQA. We believe this is due to the lack of a considerably large, annotated dataset for CQA summarization. We create CQASUMM, the first huge annotated CQA summarization dataset by filtering the 4.4 million Yahoo! Answers L6 dataset. We sample threads where the best answer can double up as a reference summary and build hundred word summaries from them. We treat other answers as candidates documents for summarization. We provide a script to generate the dataset and introduce the new task of Community Question Answering Summarization. Multi document summarization has been widely studied with news article datasets, especially in the DUC and TAC challenges using news corpora. However documents in CQA have higher variance, contradicting opinion and lesser amount of overlap. We compare the popular multi document summarization techniques and evaluate their performance on our CQA corpora. We look into the state-of-the-art and understand the cases where existing multi document summarizers (MDS) fail. We find that most MDS workflows are built for the entirely factual news corpora, whereas our corpus has a fair share of opinion based instances too. We therefore introduce OpinioSumm, a new MDS which outperforms the best baseline by 4.6% w.r.t ROUGE-1 score.

cs.CL↗

arXiv2018

Ensemble-based Overlapping Community Detection using Disjoint Community Structures

While there has been a plethora of approaches for detecting disjoint communities from real-world complex networks, some methods for detecting overlapping community structures have also been recently proposed. In this work, we argue that, instead of developing separate approaches for detecting overlapping communities, a promising alternative is to infer the overlapping communities from multiple disjoint community structures. We propose an ensemble-based approach, called EnCoD, that leverages the solutions produced by various disjoint community detection algorithms to discover the overlapping community structure. Specifically, EnCoD generates a feature vector for each vertex from the results of the base algorithms and learns which features lead to detect densely connected overlapping regions in an unsupervised way. It keeps on iterating until the likelihood of each vertex belonging to its own community maximizes. Experiments on both synthetic and several real-world networks (with known ground-truth community structures) reveal that EnCoD significantly outperforms nine state-of-the-art overlapping community detection algorithms. Finally, we show that EnCoD is generic enough to be applied to networks where the vertices are associated with explicit semantic features. To the best of our knowledge, EnCoD is the second ensemble-based overlapping community detection approach after MEDOC [1].

cs.SI↗

arXiv2018

How did the discussion go: Discourse act classification in social media conversations

We propose a novel attention based hierarchical LSTM model to classify discourse act sequences in social media conversations, aimed at mining data from online discussion using textual meanings beyond sentence level. The very uniqueness of the task is the complete categorization of possible pragmatic roles in informal textual discussions, contrary to extraction of question-answers, stance detection or sarcasm identification which are very much role specific tasks. Early attempt was made on a Reddit discussion dataset. We train our model on the same data, and present test results on two different datasets, one from Reddit and one from Facebook. Our proposed model outperformed the previous one in terms of domain independence; without using platform-dependent structural features, our hierarchical LSTM with word relevance attention mechanism achieved F1-scores of 71\% and 66\% respectively to predict discourse roles of comments in Reddit and Facebook discussions. Efficiency of recurrent and convolutional architectures in order to learn discursive representation on the same task has been presented and analyzed, with different word and comment embedding schemes. Our attention mechanism enables us to inquire into relevance ordering of text segments according to their roles in discourse. We present a human annotator experiment to unveil important observations about modeling and data annotation. Equipped with our text-based discourse identification model, we inquire into how heterogeneous non-textual features like location, time, leaning of information etc. play their roles in charaterizing online discussions on Facebook.

cs.CL↗

arXiv2018

On Good and Bad Intentions behind Anomalous Citation Patterns among Journals in Computer Sciences

Scientific journals are an important choice of publication venue for most authors. Publishing in prestigious journal plays a decisive role for authors in hiring and promotions. In last decade, citation pressure has become intact for all scientific entities more than ever before. Unethical publication practices has started to manipulate widely used performance metric such as "impact factor" for journals and citation based indices for authors. This threatens the integrity of scientific quality and takes away deserved credit of legitimate authors and their authentic publications. In this paper we extract all possible anomalous citation patterns between journals from a Computer Science bibliographic dataset which contains more than 2,500 journals. Apart from excessive self-citations, we mostly focus on finding several patterns between two or more journals such as bi-directional mutual citations, chains, triangles, mesh, cartel relationships. On a macroscopic scale, the motivation is to understand the nature of these patterns by modeling how journals mutually interact through citations. On microscopic level, we differentiate between possible intentions (good or bad) behind such patterns. We see whether such patterns prevail for long period or during any specific time duration. For abnormal citation behavior, we study the nature of sudden inflation in impact factor of journals on a time basis which may occur due to addition of irrelevant and superfluous citations in such closed pattern interaction. We also study possible influences such as abrupt increase in paper count due to the presence of self-referential papers or duplicate manuscripts, author self-citation, author co-authorship network, author-editor network, publication houses etc. The entire study is done to question the reliability of existing bibliometrics, and hence, it is an urgent need to curtail their usage or redefine them.

cs.DL↗

arXiv2018

Retweet Us, We Will Retweet You: Spotting Collusive Retweeters Involved in Blackmarket Services

Twitter has increasingly become a popular platform to share news and user opinion. A tweet is considered to be important if it receives high number of affirmative reactions from other Twitter users via Retweets. Retweet count is thus considered as a surrogate measure for positive crowd-sourced reactions - high number of retweets of a tweet aid in making its topic trending. This in turn bolsters the social reputation of the author of the tweet. Since social reputation/impact of users/tweets influences many decisions (such as promoting brands, advertisement etc.), several blackmarket syndicates have actively been engaged in producing fake retweets in a collusive manner. Users who want to boost the impact of their tweets approach the blackmarket services, and gain retweets for their own tweets by either paying money (Premium Services) or by retweeting other customers' tweets. Thus they become customers of blackmarket syndicates and engage in fake activities. Interestingly, these customers are neither bots, nor even fake users - they are usually normal human beings; they express a mix of organic and inorganic retweeting activities, and there is no synchronicity across their behaviors. In this paper, we make a first attempt to investigate such blackmarket customers engaged in producing fake retweets. We collected and annotated a novel dataset comprising of customers of many blackmarket services and show how their social behavior differs from genuine users. We then use state-of-the-art supervised models to detect three types of customers (bots, promotional, normal) and genuine users. We achieve a Macro F1-score of 0.87 with SVM, outperforming four other baselines significantly. We further design a browser extension, SCoRe which, given the link of a tweet, spots its fake retweeters in real-time. We also collected users' feedback on the performance of SCoRe and obtained 85% accuracy.

cs.SI↗

arXiv2018

MMGAN: Manifold Matching Generative Adversarial Network

It is well-known that GANs are difficult to train, and several different techniques have been proposed in order to stabilize their training. In this paper, we propose a novel training method called manifold-matching, and a new GAN model called manifold-matching GAN (MMGAN). MMGAN finds two manifolds representing the vector representations of real and fake images. If these two manifolds match, it means that real and fake images are statistically identical. To assist the manifold-matching task, we also use i) kernel tricks to find better manifold structures, ii) moving-averaged manifolds across mini-batches, and iii) a regularizer based on correlation matrix to suppress mode collapse. We conduct in-depth experiments with three image datasets and compare with several state-of-the-art GAN models. 32.4% of images generated by the proposed MMGAN are recognized as fake images during our user study (16% enhancement compared to other state-of-the-art model). MMGAN achieved an unsupervised inception score of 7.8 for CIFAR-10.

cs.LG↗

arXiv2018

DyPerm: Maximizing Permanence for Dynamic Community Detection

In this paper, we propose DyPerm, the first dynamic community detection method which optimizes a novel community scoring metric, called permanence. DyPerm incrementally modifies the community structure by updating those communities where the editing of nodes and edges has been performed, keeping the rest of the network unchanged. We present strong theoretical guarantees to show how/why mere updates on the existing community structure leads to permanence maximization in dynamic networks, which in turn decreases the computational complexity drastically. Experiments on both synthetic and six real-world networks with given ground-truth community structure show that DyPerm achieves (on average) 35% gain in accuracy (based on NMI) compared to the best method among four baseline methods. DyPerm also turns out to be 15 times faster than its static counterpart.

cs.SI↗

arXiv2018

Collective Classification of Spam Campaigners on Twitter: A Hierarchical Meta-Path Based Approach

Cybercriminals have leveraged the popularity of a large user base available on Online Social Networks to spread spam campaigns by propagating phishing URLs, attaching malicious contents, etc. However, another kind of spam attacks using phone numbers has recently become prevalent on OSNs, where spammers advertise phone numbers to attract users' attention and convince them to make a call to these phone numbers. The dynamics of phone number based spam is different from URL-based spam due to an inherent trust associated with a phone number. While previous work has proposed strategies to mitigate URL-based spam attacks, phone number based spam attacks have received less attention. In this paper, we aim to detect spammers that use phone numbers to promote campaigns on Twitter. To this end, we collected information about 3,370 campaigns spread by 670,251 users. We model the Twitter dataset as a heterogeneous network by leveraging various interconnections between different types of nodes present in the dataset. In particular, we make the following contributions: (i) We propose a simple yet effective metric, called Hierarchical Meta-Path Score (HMPS) to measure the proximity of an unknown user to the other known pool of spammers. (ii) We design a feedback-based active learning strategy and show that it significantly outperforms three state-of-the-art baselines for the task of spam detection. Our method achieves 6.9% and 67.3% higher F1-score and AUC, respectively compared to the best baseline method. (iii) To overcome the problem of less training instances for supervised learning, we show that our proposed feedback strategy achieves 25.6% and 46% higher F1-score and AUC respectively than other oversampling strategies. Finally, we perform a case study to show how our method is capable of detecting those users as spammers who have not been suspended by Twitter (and other baselines) yet.

cs.SI↗

arXiv2018

ComPAS: Community Preserving Sampling for Streaming Graphs

In the era of big data, graph sampling is indispensable in many settings. Existing sampling methods are mostly designed for static graphs, and aim to preserve basic structural properties of the original graph (such as degree distribution, clustering coefficient etc.) in the sample. We argue that for any sampling method it is impossible to produce an universal representative sample which can preserve all the properties of the original graph; rather sampling should be application specific (such as preserving hubs - needed for information diffusion). Here we consider community detection as an application scenario. We propose ComPAS, a novel sampling strategy that unlike previous methods, is not only designed for streaming graphs (which is a more realistic representation of a real-world scenario) but also preserves the community structure of the original graph in the sample. Empirical results on both synthetic and different real-world graphs show that ComPAS is the best to preserve the underlying community structure with average performance reaching 73.2% of the most informed algorithm for static graphs.

cs.SI↗

arXiv2018

Mining Twitter Conversations around E-commerce Promotional Events

With Social Media platforms establishing themselves as the de facto destinations for their customers views and opinions, brands around the World are investing heavily on invigorating their customer connects by utilizing such platforms to their fullest. In this paper, we develop a novel technique for mining conversations in Twitter by weaving together all conversations around an event into one unified graph (Conversation Graph, henceforth). The structure of the Conversation Graph emerges as a variant of the BOWTIE structure (dubbed ASKEWBOWTIE henceforth) as a result of the complex communication patterns amongst these players. Finally, we investigate the structural properties of the ASKEWBOWTIE structure to understand the configuration of the components and their temporal evolution.

cs.SI↗

arXiv2017

Ensemble-Based Discovery of Disjoint, Overlapping and Fuzzy Community Structures in Networks

Though much work has been done on ensemble clustering in data mining, the application of ensemble methods to community detection in networks is in its infancy. In this paper, we propose two ensemble methods: ENDISCO and MEDOC. ENDISCO performs disjoint community detection. In contrast, MEDOC performs disjoint, overlapping, and fuzzy community detection and represents the first ever ensemble method for fuzzy and overlapping community detection. We run extensive experiments with both algorithms against both synthetic and several real-world datasets for which community structures are known. We show that ENDISCO and MEDOC both beat the best-known existing standalone community detection algorithms (though we emphasize that they leverage them). In the case of disjoint community detection, we show that both ENDISCO and MEDOC beat an existing ensemble community detection algorithm both in terms of multiple accuracy measures and run-time. We further show that our ensemble algorithms can help explore core-periphery structure of network communities, identify stable communities in dynamic networks and help solve the "degeneracy of solutions" problem, generating robust results.

cs.SI↗

arXiv2017

Role of Interdisciplinarity in Computer Sciences: Quantification, Impact and Life Trajectory

The tremendous advances in computer science in the last few decades have provided the platform to address and solve complex problems using interdisciplinary research. In this paper, we investigate how the extent of interdisciplinarity in computer science domain (which is further divided into 24 research fields) has changed over the last 50 years. To this end, we collect a massive bibliographic dataset with rich metadata information. We start with quantifying interdisciplinarity of a field in terms of the diversity of topics and citations. We then analyze the effect of interdisciplinary research on the scientific impact of individual fields and observe that highly disciplinary and highly interdisciplinary papers in general have a low scientific impact; remarkably those that are able to strike a balance between the two extremes eventually land up having the highest impact. Further, we study the reciprocity among fields through citation interactions and notice that links from one field to related and citation-intensive fields (fields producing large number of citations) are reciprocated heavily. A systematic analysis of the citation interactions reveals the life trajectory of a research field, which generally undergoes three phases -- a growing phase, a matured phase and an interdisciplinary phase. The combination of metrics and empirical observations presented here provides general benchmarks for future studies of interdisciplinary research activities in other domains of science.

cs.DL↗

arXiv2017

EC3: Combining Clustering and Classification for Ensemble Learning

Classification and clustering algorithms have been proved to be successful individually in different contexts. Both of them have their own advantages and limitations. For instance, although classification algorithms are more powerful than clustering methods in predicting class labels of objects, they do not perform well when there is a lack of sufficient manually labeled reliable data. On the other hand, although clustering algorithms do not produce label information for objects, they provide supplementary constraints (e.g., if two objects are clustered together, it is more likely that the same label is assigned to both of them) that one can leverage for label prediction of a set of unknown objects. Therefore, systematic utilization of both these types of algorithms together can lead to better prediction performance. In this paper, We propose a novel algorithm, called EC3 that merges classification and clustering together in order to support both binary and multi-class classification. EC3 is based on a principled combination of multiple classification and multiple clustering methods using an optimization function. We theoretically show the convexity and optimality of the problem and solve it by block coordinate descent method. We additionally propose iEC3, a variant of EC3 that handles imbalanced training data. We perform an extensive experimental analysis by comparing EC3 and iEC3 with 14 baseline methods (7 well-known standalone classifiers, 5 ensemble classifiers, and 2 existing methods that merge classification and clustering) on 13 standard benchmark datasets. We show that our methods outperform other baselines for every single dataset, achieving at most 10% higher AUC. Moreover our methods are faster (1.21 times faster than the best baseline), more resilient to noise and class imbalance than the best baseline method.

cs.LG↗