arXiv subjects

Vikram Goyal

Publications and source records attributed to Vikram Goyal.

31 records · Page 2Linked to original sources

arXiv2022

Auxiliary Task Guided Interactive Attention Model for Question Difficulty Prediction

Online learning platforms conduct exams to evaluate the learners in a monotonous way, where the questions in the database may be classified into Bloom's Taxonomy as varying levels in complexity from basic knowledge to advanced evaluation. The questions asked in these exams to all learners are very much static. It becomes important to ask new questions with different difficulty levels to each learner to provide a personalized learning experience. In this paper, we propose a multi-task method with an interactive attention mechanism, Qdiff, for jointly predicting Bloom's Taxonomy and difficulty levels of academic questions. We model the interaction between the predicted bloom taxonomy representations and the input representations using an attention mechanism to aid in difficulty prediction. The proposed learning method would help learn representations that capture the relationship between Bloom's taxonomy and difficulty labels. The proposed multi-task method learns a good input representation by leveraging the relationship between the related tasks and can be used in similar settings where the tasks are related. The results demonstrate that the proposed method performs better than training only on difficulty prediction. However, Bloom's labels may not always be given for some datasets. Hence we soft label another dataset with a model fine-tuned to predict Bloom's labels to demonstrate the applicability of our method to datasets with only difficulty labels.

cs.CY↗

arXiv2022

Topic Aware Contextualized Embeddings for High Quality Phrase Extraction

Keyphrase extraction from a given document is the task of automatically extracting salient phrases that best describe the document. This paper proposes a novel unsupervised graph-based ranking method to extract high-quality phrases from a given document. We obtain the contextualized embeddings from pre-trained language models enriched with topic vectors from Latent Dirichlet Allocation (LDA) to represent the candidate phrases and the document. We introduce a scoring mechanism for the phrases using the information obtained from contextualized embeddings and the topic vectors. The salient phrases are extracted using a ranking algorithm on an undirected graph constructed for the given document. In the undirected graph, the nodes represent the phrases, and the edges between the phrases represent the semantic relatedness between them, weighted by a score obtained from the scoring mechanism. To demonstrate the efficacy of our proposed method, we perform several experiments on open source datasets in the science domain and observe that our novel method outperforms existing unsupervised embedding based keyphrase extraction methods. For instance, on the SemEval2017 dataset, our method advances the F1 score from 0.2195 (EmbedRank) to 0.2819 at the top 10 extracted keyphrases. Several variants of the proposed algorithm are investigated to determine their effect on the quality of keyphrases. We further demonstrate the ability of our proposed method to collect additional high-quality keyphrases that are not present in the document from external knowledge bases like Wikipedia for enriching the document with newly discovered keyphrases. We evaluate this step on a collection of annotated documents. The F1-score at the top 10 expanded keyphrases is 0.60, indicating that our algorithm can also be used for 'concept' expansion using external knowledge.

cs.IR↗

arXiv2021

See, Hear, Read: Leveraging Multimodality with Guided Attention for Abstractive Text Summarization

In recent years, abstractive text summarization with multimodal inputs has started drawing attention due to its ability to accumulate information from different source modalities and generate a fluent textual summary. However, existing methods use short videos as the visual modality and short summary as the ground-truth, therefore, perform poorly on lengthy videos and long ground-truth summary. Additionally, there exists no benchmark dataset to generalize this task on videos of varying lengths. In this paper, we introduce AVIATE, the first large-scale dataset for abstractive text summarization with videos of diverse duration, compiled from presentations in well-known academic conferences like NDSS, ICML, NeurIPS, etc. We use the abstract of corresponding research papers as the reference summaries, which ensure adequate quality and uniformity of the ground-truth. We then propose FLORAL, a factorized multi-modal Transformer based decoder-only language model, which inherently captures the intra-modal and inter-modal dynamics within various input modalities for the text summarization task. FLORAL utilizes an increasing number of self-attentions to capture multimodality and performs significantly better than traditional encoder-decoder based networks. Extensive experiments illustrate that FLORAL achieves significant improvement over the baselines in both qualitative and quantitative evaluations on the existing How2 dataset for short videos and newly introduced AVIATE dataset for videos with diverse duration, beating the best baseline on the two datasets by $1.39$ and $2.74$ ROUGE-L points respectively.

cs.LG↗

arXiv2021

TagRec: Automated Tagging of Questions with Hierarchical Learning Taxonomy

Online educational platforms organize academic questions based on a hierarchical learning taxonomy (subject-chapter-topic). Automatically tagging new questions with existing taxonomy will help organize these questions into different classes of hierarchical taxonomy so that they can be searched based on the facets like chapter. This task can be formulated as a flat multi-class classification problem. Usually, flat classification based methods ignore the semantic relatedness between the terms in the hierarchical taxonomy and the questions. Some traditional methods also suffer from the class imbalance issues as they consider only the leaf nodes ignoring the hierarchy. Hence, we formulate the problem as a similarity-based retrieval task where we optimize the semantic relatedness between the taxonomy and the questions. We demonstrate that our method helps to handle the unseen labels and hence can be used for taxonomy tagging in the wild. In this method, we augment the question with its corresponding answer to capture more semantic information and then align the question-answer pair's contextualized embedding with the corresponding label (taxonomy) vector representations. The representations are aligned by fine-tuning a transformer based model with a loss function that is a combination of the cosine similarity and hinge rank loss. The loss function maximizes the similarity between the question-answer pair and the correct label representations and minimizes the similarity to unrelated labels. Finally, we perform experiments on two real-world datasets. We show that the proposed learning method outperforms representations learned using the multi-class classification method and other state of the art methods by 6% as measured by Recall@k. We also demonstrate the performance of the proposed method on unseen but related learning content like the learning objectives without re-training the network.

cs.CL↗

arXiv2021

THUE: Discovering Top-K High Utility Episodes

Episode discovery from an event is a popular framework for data mining tasks and has many real-world applications. An episode is a partially ordered set of objects (e.g., item, node), and each object is associated with an event type. This episode can also be considered as a complex event sub-sequence. High-utility episode mining is an interesting utility-driven mining task in the real world. Traditional episode mining algorithms, by setting a threshold, usually return a huge episode that is neither intuitive nor saves time. In general, finding a suitable threshold in a pattern-mining algorithm is a trivial and time-consuming task. In this paper, we propose a novel algorithm, called Top-K High Utility Episode (THUE) mining within the complex event sequence, which redefines the previous mining task by obtaining the K highest episodes. We introduce several threshold-raising strategies and optimize the episode-weighted utilization upper bounds to speed up the mining process and effectively reduce the memory cost. Finally, the experimental results on both real-life and synthetic datasets reveal that the THUE algorithm can offer six to eight orders of magnitude running time performance improvement over the state-of-the-art algorithm and has low memory consumption.

cs.DB↗

arXiv2020

Hate is the New Infodemic: A Topic-aware Modeling of Hate Speech Diffusion on Twitter

Online hate speech, particularly over microblogging platforms like Twitter, has emerged as arguably the most severe issue of the past decade. Several countries have reported a steep rise in hate crimes infuriated by malicious hate campaigns. While the detection of hate speech is one of the emerging research areas, the generation and spread of topic-dependent hate in the information network remain under-explored. In this work, we focus on exploring user behaviour, which triggers the genesis of hate speech on Twitter and how it diffuses via retweets. We crawl a large-scale dataset of tweets, retweets, user activity history, and follower networks, comprising over 161 million tweets from more than $41$ million unique users. We also collect over 600k contemporary news articles published online. We characterize different signals of information that govern these dynamics. Our analyses differentiate the diffusion dynamics in the presence of hate from usual information diffusion. This motivates us to formulate the modelling problem in a topic-aware setting with real-world knowledge. For predicting the initiation of hate speech for any given hashtag, we propose multiple feature-rich models, with the best performing one achieving a macro F1 score of 0.65. Meanwhile, to predict the retweet dynamics on Twitter, we propose RETINA, a novel neural architecture that incorporates exogenous influence using scaled dot-product attention. RETINA achieves a macro F1-score of 0.85, outperforming multiple state-of-the-art models. Our analysis reveals the superlative power of RETINA to predict the retweet dynamics of hateful content compared to the existing diffusion models.

cs.SI↗

arXiv2020

High-utility itemset mining for subadditive monotone utility functions

High-utility Itemset Mining (HUIM) finds itemsets from a transaction database with utility no less than a user-defined threshold where the utility of an itemset is defined as the sum of the item-wise utilities. In this paper, we generalize this notion to utility functions that need not be a simple sum of individual utilities. In particular, we study generalized utility functions that are subadditive and monotone (SM). We also describe a novel function that allows us to include external information in the form of a relationship graph for computing utility. Next, we focus on algorithms for HUIM problems with SM utility functions. We note that the existing HUIM algorithms use upper-bounds like "Transaction Weighted Utility" and "Exact-Utility, Remaining-Utility" for efficient search-space exploration. We derive analogous and tighter upper-bounds for SM utility functions. We design a novel inverted-list data structure called SMI-list and a new algorithm called SM-Miner to mine HUIs for SM functions. We explain how existing tree-based and projection-based HUIM algorithms can be adapted using these bounds. We experimentally compare adaptations of some of the latest HUIM algorithms and point out some caveats that should be kept in mind while handling utility functions that allow integration of domain knowledge with a transaction database.

cs.DB↗

arXiv2020

Interpretability of Blackbox Machine Learning Models through Dataview Extraction and Shadow Model creation

Deep learning models trained using massive amounts of data tend to capture one view of the data and its associated mapping. Different deep learning models built on the same training data may capture different views of the data based on the underlying techniques used. For explaining the decisions arrived by blackbox deep learning models, we argue that it is essential to reproduce that model's view of the training data faithfully. This faithful reproduction can then be used for explanation generation. We investigate two methods for data view extraction: hill-climbing approach and a GAN-driven approach. We then use this synthesized data for creating shadow models for explanation generation: Decision-Tree model and Formal Concept Analysis based model. We evaluate these approaches on a Blackbox model trained on public datasets and show its usefulness in explanation generation.

cs.AI↗

arXiv2019

Mining Top-k Trajectory-Patterns from Anonymized Data

The ubiquity of GPS enabled devices result into the generation of an enormous amount of user movement data consisting of a sequence of locations together with activities performed at those locations. Such data, commonly known as {\it activity-trajectory data}, can be analysed to find common user movements and preferred activities, which will have tremendous business value. However, various countries have now introduced data privacy regulations that make it mandatory to anonymize any data before releasing it. This makes it difficult to look for patterns as the existing mining techniques may not be directly applicable over anonymized data. User locations in an activity-trajectory dataset are anonymized to regions of different shapes and sizes making them uncomparable; therefore, it is unclear how to define suitable patterns over those regions. In this paper, we propose a top-k pattern mining technique called TopKMintra that employs a pre-processing step to transform anonymized activity-trajectory into an intermediate representation that address the above problem. Unlike usual sequential data, activity-trajectory data is 2-dimensional that can lead to generation of duplicate patterns. To handle this, TopKMintra restricts arbitrary extensions in the two dimensions by imposing an order over the search space; this also helps in addressing the common problem of updating the threshold in top-k pattern mining algorithms during various stages. We perform extensive experiments on real datasets to demonstrate the efficiency and the effectiveness of TopKMintra. Our results show that even after anonymization, certain patterns may remain in a dataset and those could be mined efficiently.

cs.DS↗

arXiv2019

A one-phase tree-based algorithm for mining high-utility itemsets from a transaction database

High-utility itemset mining finds itemsets from a transaction database with utility no less than a fixed user-defined threshold. The utility of an itemset is defined as the sum of the utilities of its item. Several algorithms were proposed to mine high-utility itemsets. However, no state-of-the-art algorithm performs consistently good across dense and sparse datasets. In this paper, we propose a novel data structure called Utility-Tree, and a tree-based algorithm called UT-Miner that mines high-utility itemsets in one-phase only without generating any candidates and uses a lightweight construction method to reduce the cost of creating projected databases during the search space exploration. The transaction information is stored compactly with every node of the Utility-Tree, and the information is computed efficiently during the recursive invocation of the algorithm. Experimental results on several real-life dense and sparse datasets reveal that UT-Miner is among the top-performing efficient algorithms across different datasets.

cs.DS↗

arXiv2019

DiffQue: Estimating Relative Difficulty of Questions in Community Question Answering Services

Automatic estimation of relative difficulty of a pair of questions is an important and challenging problem in community question answering (CQA) services. There are limited studies which addressed this problem. Past studies mostly leveraged expertise of users answering the questions and barely considered other properties of CQA services such as metadata of users and posts, temporal information and textual content. In this paper, we propose DiffQue, a novel system that maps this problem to a network-aided edge directionality prediction problem. DiffQue starts by constructing a novel network structure that captures different notions of difficulties among a pair of questions. It then measures the relative difficulty of two questions by predicting the direction of a (virtual) edge connecting these two questions in the network. It leverages features extracted from the network structure, metadata of users/posts and textual description of questions and answers. Experiments on datasets obtained from two CQA sites (further divided into four datasets) with human annotated ground-truth show that DiffQue outperforms four state-of-the-art methods by a significant margin (28.77% higher F1 score and 28.72% higher AUC than the best baseline). As opposed to the other baselines, (i) DiffQue appropriately responds to the training noise, (ii) DiffQue is capable of adapting multiple domains (CQA datasets), and (iii) DiffQue can efficiently handle 'cold start' problem which may arise due to the lack of information for newly posted questions or newly arrived users.

cs.LG↗

arXiv2018

Extracting Fairness Policies from Legal Documents

Machine Learning community is recently exploring the implications of bias and fairness with respect to the AI applications. The definition of fairness for such applications varies based on their domain of application. The policies governing the use of such machine learning system in a given context are defined by the constitutional laws of nations and regulatory policies enforced by the organizations that are involved in the usage. Fairness related laws and policies are often spread across the large documents like constitution, agreements, and organizational regulations. These legal documents have long complex sentences in order to achieve rigorousness and robustness. Automatic extraction of fairness policies, or in general, any specific kind of policies from large legal corpus can be very useful for the study of bias and fairness in the context of AI applications. We attempted to automatically extract fairness policies from publicly available law documents using two approaches based on semantic relatedness. The experiments reveal how classical Wordnet-based similarity and vector-based similarity differ in addressing this task. We have shown that similarity based on word vectors beats the classical approach with a large margin, whereas other vector representations of senses and sentences fail to even match the classical baseline. Further, we have presented thorough error analysis and reasoning to explain the results with appropriate examples from the dataset for deeper insights.

cs.LG↗

arXiv2015

Efficient Reverse k Nearest Neighbor evaluation for hierarchical index

"Reverse Nearest Neighbor" query finds applications in decision support systems, profile-based marketing, emergency services etc. In this paper, we point out a few flaws in the branch and bound algorithms proposed earlier for computing monochromatic RkNN queries over data points stored in hierarchical index. We give suitable counter examples to validate our claims and propose a correct algorithm for the corresponding problem. We show that our algorithm is correct by identifying necessary conditions behind correctness of algorithms for this problem.

cs.DS↗