arXiv subjects

Roshni Chakraborty

Publications and source records attributed to Roshni Chakraborty.

15 recordsLinked to original sources

arXiv2026

When Features Become Instances: Inverted Contrastive Learning for Unsupervised Feature Selection

Unsupervised feature selection seeks a compact subset of informative features without access to class labels, making feature utility difficult to define. Existing UFS methods therefore rely on indirect structural criteria, such as similarity preservation, locality, sparsity, cluster geometry, or reconstruction quality. In this paper, we instead study UFS through representation consistency and propose Inverted Contrastive Learning for Unsupervised Feature Selection (ICLFS), a feature-wise contrastive framework that reformulates UFS as a representation learning problem over features rather than samples. ICLFS first inverts the data matrix so that each feature is represented by its sample-profile vector, then constructs multiple masked positive views together with a shuffled negative view, and learns projector-space representations that remain consistent across these structured perturbations under an InfoNCE-based objective. Motivated by recent findings that cosine-based and InfoNCE-based training affect embedding norms, we use projector-space embedding magnitude as the saliency signal for ranking features. The resulting norm-based ranking is subsequently refined through Laplacian-Gated Ranking Correction, which suppresses locally redundant candidates while preserving salient ones. Extensive experiments on 12 benchmark datasets show that ICLFS achieves the best clustering accuracy on 10 datasets against both classical and neural baselines under the standard clustering-based UFS evaluation protocol, while remaining competitive on the other two. These results show that feature-wise contrastive representation consistency provides a strong and effective alternative to neighborhood, cluster, and reconstruction-based UFS formulations.

cs.AI↗

arXiv2026

MADS: A Multiview Acoustic Descriptor Set Beyond Standard Spectral Summaries

Dominant audio classification pipelines rely either on compact handcrafted summaries or on fixed time-frequency frontends such as log-mel representations prior to deep modeling. While highly successful, these representations do not explicitly expose the physical dynamics of the underlying sound-generating event. We introduce MADS (Multi-view Acoustic Descriptor Set), a compact 19-dimensional physics-informed descriptor set de- signed to capture complementary spectral, temporal, mechanical, and stochastic structure in audio signals. Rather than treating sound only as a spectral pattern, MADS encodes properties related to excitation, damping, periodicity, impulsiveness, and structural consistency within a unified multi-view representation. We evaluate MADS using standard classical machine learning models on ESC-10, ESC-50, and MSoS, and compare it against two conventional handcrafted baselines: a compact 26D MFCC- based baseline and an expanded 38D spectral-summary baseline. Across ESC-10 and ESC-50, MADS achieves the strongest peak results overall, reaching 81.00% and 52.78%, respectively, while using roughly half the dimensionality of the 38D baseline. On MSoS, MADS again delivers the strongest top-end performance, reaching 67.48%. These results establish MADS not merely as a competitive standalone descriptor set, but as the foundational descriptor layer of a broader acoustically grounded representation program for future frame-level and deep-learning-compatible audio modeling.

cs.SD↗

arXiv2026

EquiSumm : A Gender Bias-Aware Framework for Inclusive Tweet Summarization

While social media platforms, such as Twitter, provide a medium for large-scale opinion sharing during news events, it is manually impossible for individuals or media agencies to process the vast volume of content to identify key viewpoints. In order to resolve this, several automatic summarization techniques have been proposed to condense large collections of tweets into concise and informative summaries. However, these algorithms do not explicitly consider demographic fairness. Several existing research works have developed automated summarization approaches that can provide a holistic overview of the key aspects and major opinions shared on social media platforms related to a news event. However, these approaches do not explicitly consider different forms of demographic representation, such as gender, which can lead to biased summary representation. In this paper, we propose EquiSumm, which considers the gender aspect of the shared opinion to generate a summary, and our experimental analysis on two major datasets indicates the performance effectiveness with respect to existing research works.

cs.CL↗

arXiv2026

IMSAHLO: Integrating Multi-Scale Attention and Hybrid Loss Optimization Framework for Robust Neuronal Brain Cell Segmentation

Accurate segmentation of neuronal cells in fluorescence microscopy is a fundamental task for quantitative analysis in computational neuroscience. However, it is significantly impeded by challenges such as the coexistence of densely packed and sparsely distributed cells, complex overlapping morphologies, and severe class imbalance. Conventional deep learning models often fail to preserve fine topological details or accurately delineate boundaries under these conditions. To address these limitations, we propose a novel deep learning framework, IMSAHLO (Integrating Multi-Scale Attention and Hybrid Loss Optimization), for robust and adaptive neuronal segmentation. The core of our model features Multi-Scale Dense Blocks (MSDBs) to capture features at various receptive fields, effectively handling variations in cell density, and a Hierarchical Attention (HA) mechanism that adaptively focuses on salient morphological features to preserve Region of Interest (ROI) boundary details. Furthermore, we introduce a novel hybrid loss function synergistically combining Tversky and Focal loss to combat class imbalance, alongside a topology-aware Centerline Dice (clDice) loss and a Contour-Weighted Boundary loss to ensure topological continuity and precise separation of adjacent cells. Large-scale experiments on the public Fluorescent Neuronal Cells (FNC) dataset demonstrate that our framework outperforms state-of-the-art architectures, achieving precision of 81.4%, macro F1 score of 82.7%, micro F1 score of 83.3%, and balanced accuracy of 99.5% on difficult dense and sparse cases. Ablation studies validate the synergistic benefits of multi-scale attention and hybrid loss terms. This work establishes a foundation for generalizable segmentation models applicable to a wide range of biomedical imaging modalities, pushing AI-assisted analysis toward high-throughput neurobiological pipelines.

cs.CV↗

arXiv2024

Who Checks the Checkers? Exploring Source Credibility in Twitter's Community Notes

In recent years, the proliferation of misinformation on social media platforms has become a significant concern. Initially designed for sharing information and fostering social connections, platforms like Twitter (now rebranded as X) have also unfortunately become conduits for spreading misinformation. To mitigate this, these platforms have implemented various mechanisms, including the recent suggestion to use crowd-sourced non-expert fact-checkers to enhance the scalability and efficiency of content vetting. An example of this is the introduction of Community Notes on Twitter. While previous research has extensively explored various aspects of Twitter tweets, such as information diffusion, sentiment analytics and opinion summarization, there has been a limited focus on the specific feature of Twitter Community Notes, despite its potential role in crowd-sourced fact-checking. Prior research on Twitter Community Notes has involved empirical analysis of the feature's dataset and comparative studies that also include other methods like expert fact-checking. Distinguishing itself from prior works, our study covers a multi-faceted analysis of sources and audience perception within Community Notes. We find that the majority of cited sources are news outlets that are left-leaning and are of high factuality, pointing to a potential bias in the platform's community fact-checking. Left biased and low factuality sources validate tweets more, while Center sources are used more often to refute tweet content. Additionally, source factuality significantly influences public agreement and helpfulness of the notes, highlighting the effectiveness of the Community Notes Ranking algorithm. These findings showcase the impact and biases inherent in community-based fact-checking initiatives.

cs.CY↗

arXiv2024

ATSumm: Auxiliary information enhanced approach for abstractive disaster Tweet Summarization with sparse training data

The abundance of situational information on Twitter poses a challenge for users to manually discern vital and relevant information during disasters. A concise and human-interpretable overview of this information helps decision-makers in implementing efficient and quick disaster response. Existing abstractive summarization approaches can be categorized as sentence-based or key-phrase-based approaches. This paper focuses on sentence-based approach, which is typically implemented as a dual-phase procedure in literature. The initial phase, known as the extractive phase, involves identifying the most relevant tweets. The subsequent phase, referred to as the abstractive phase, entails generating a more human-interpretable summary. In this study, we adopt the methodology from prior research for the extractive phase. For the abstractive phase of summarization, most existing approaches employ deep learning-based frameworks, which can either be pre-trained or require training from scratch. However, to achieve the appropriate level of performance, it is imperative to have substantial training data for both methods, which is not readily available. This work presents an Abstractive Tweet Summarizer (ATSumm) that effectively addresses the issue of data sparsity by using auxiliary information. We introduced the Auxiliary Pointer Generator Network (AuxPGN) model, which utilizes a unique attention mechanism called Key-phrase attention. This attention mechanism incorporates auxiliary information in the form of key-phrases and their corresponding importance scores from the input tweets. We evaluate the proposed approach by comparing it with 10 state-of-the-art approaches across 13 disaster datasets. The evaluation results indicate that ATSumm achieves superior performance compared to state-of-the-art approaches, with improvement of 4-80% in ROUGE-N F1-score.

cs.CL↗

arXiv2024

ADSumm: Annotated Ground-truth Summary Datasets for Disaster Tweet Summarization

Online social media platforms, such as Twitter, provide valuable information during disaster events. Existing tweet disaster summarization approaches provide a summary of these events to aid government agencies, humanitarian organizations, etc., to ensure effective disaster response. In the literature, there are two types of approaches for disaster summarization, namely, supervised and unsupervised approaches. Although supervised approaches are typically more effective, they necessitate a sizable number of disaster event summaries for testing and training. However, there is a lack of good number of disaster summary datasets for training and evaluation. This motivates us to add more datasets to make supervised learning approaches more efficient. In this paper, we present ADSumm, which adds annotated ground-truth summaries for eight disaster events which consist of both natural and man-made disaster events belonging to seven different countries. Our experimental analysis shows that the newly added datasets improve the performance of the supervised summarization approaches by 8-28% in terms of ROUGE-N F1-score. Moreover, in newly annotated dataset, we have added a category label for each input tweet which helps to ensure good coverage from different categories in summary. Additionally, we have added two other features relevance label and key-phrase, which provide information about the quality of a tweet and explanation about the inclusion of the tweet into summary, respectively. For ground-truth summary creation, we provide the annotation procedure adapted in detail, which has not been described in existing literature. Experimental analysis shows the quality of ground-truth summary is very good with Coverage, Relevance and Diversity.

cs.CL↗

arXiv2023

PORTRAIT: a hybrid aPproach tO cReate extractive ground-TRuth summAry for dIsaster evenT

Disaster summarization approaches provide an overview of the important information posted during disaster events on social media platforms, such as, Twitter. However, the type of information posted significantly varies across disasters depending on several factors like the location, type, severity, etc. Verification of the effectiveness of disaster summarization approaches still suffer due to the lack of availability of good spectrum of datasets along with the ground-truth summary. Existing approaches for ground-truth summary generation (ground-truth for extractive summarization) relies on the wisdom and intuition of the annotators. Annotators are provided with a complete set of input tweets from which a subset of tweets is selected by the annotators for the summary. This process requires immense human effort and significant time. Additionally, this intuition-based selection of the tweets might lead to a high variance in summaries generated across annotators. Therefore, to handle these challenges, we propose a hybrid (semi-automated) approach (PORTRAIT) where we partly automate the ground-truth summary generation procedure. This approach reduces the effort and time of the annotators while ensuring the quality of the created ground-truth summary. We validate the effectiveness of PORTRAIT on 5 disaster events through quantitative and qualitative comparisons of ground-truth summaries generated by existing intuitive approaches, a semi-automated approach, and PORTRAIT. We prepare and release the ground-truth summaries for 5 disaster events which consist of both natural and man-made disaster events belonging to 4 different countries. Finally, we provide a study about the performance of various state-of-the-art summarization approaches on the ground-truth summaries generated by PORTRAIT using ROUGE-N F1-scores.

cs.CL↗

arXiv2023

IKDSumm: Incorporating Key-phrases into BERT for extractive Disaster Tweet Summarization

Online social media platforms, such as Twitter, are one of the most valuable sources of information during disaster events. Therefore, humanitarian organizations, government agencies, and volunteers rely on a summary of this information, i.e., tweets, for effective disaster management. Although there are several existing supervised and unsupervised approaches for automated tweet summary approaches, these approaches either require extensive labeled information or do not incorporate specific domain knowledge of disasters. Additionally, the most recent approaches to disaster summarization have proposed BERT-based models to enhance the summary quality. However, for further improved performance, we introduce the utilization of domain-specific knowledge without any human efforts to understand the importance (salience) of a tweet which further aids in summary creation and improves summary quality. In this paper, we propose a disaster-specific tweet summarization framework, IKDSumm, which initially identifies the crucial and important information from each tweet related to a disaster through key-phrases of that tweet. We identify these key-phrases by utilizing the domain knowledge (using existing ontology) of disasters without any human intervention. Further, we utilize these key-phrases to automatically generate a summary of the tweets. Therefore, given tweets related to a disaster, IKDSumm ensures fulfillment of the summarization key objectives, such as information coverage, relevance, and diversity in summary without any human intervention. We evaluate the performance of IKDSumm with 8 state-of-the-art techniques on 12 disaster datasets. The evaluation results show that IKDSumm outperforms existing techniques by approximately 2-79% in terms of ROUGE-N F1-score.

cs.CL↗

arXiv2022

OntoDSumm : Ontology based Tweet Summarization for Disaster Events

The huge popularity of social media platforms like Twitter attracts a large fraction of users to share real-time information and short situational messages during disasters. A summary of these tweets is required by the government organizations, agencies, and volunteers for efficient and quick disaster response. However, the huge influx of tweets makes it difficult to manually get a precise overview of ongoing events. To handle this challenge, several tweet summarization approaches have been proposed. In most of the existing literature, tweet summarization is broken into a two-step process where in the first step, it categorizes tweets, and in the second step, it chooses representative tweets from each category. There are both supervised as well as unsupervised approaches found in literature to solve the problem of first step. Supervised approaches requires huge amount of labelled data which incurs cost as well as time. On the other hand, unsupervised approaches could not clusters tweet properly due to the overlapping keywords, vocabulary size, lack of understanding of semantic meaning etc. While, for the second step of summarization, existing approaches applied different ranking methods where those ranking methods are very generic which fail to compute proper importance of a tweet respect to a disaster. Both the problems can be handled far better with proper domain knowledge. In this paper, we exploited already existing domain knowledge by the means of ontology in both the steps and proposed a novel disaster summarization method OntoDSumm. We evaluate this proposed method with 4 state-of-the-art methods using 10 disaster datasets. Evaluation results reveal that OntoDSumm outperforms existing methods by approximately 2-66% in terms of ROUGE-1 F1 score.

cs.SI↗

arXiv2022

EnDSUM: Entropy and Diversity based Disaster Tweet Summarization

The huge amount of information shared in Twitter during disaster events are utilized by government agencies and humanitarian organizations to ensure quick crisis response and provide situational updates. However, the huge number of tweets posted makes manual identification of the relevant tweets impossible. To address the information overload, there is a need to automatically generate summary of all the tweets which can highlight the important aspects of the disaster. In this paper, we propose an entropy and diversity based summarizer, termed as EnDSUM, specifically for disaster tweet summarization. Our comprehensive analysis on 6 datasets indicates the effectiveness of EnDSUM and additionally, highlights the scope of improvement of EnDSUM.

cs.SI↗

arXiv2022

Finding Representative Sampling Subsets in Sensor Graphs using Time Series Similarities

With the increasing use of IoT-enabled sensors, it is important to have effective methods for querying the sensors. For example, in a dense network of battery-driven temperature sensors, it is often possible to query (sample) just a subset of the sensors at any given time, since the values of the non-sampled sensors can be estimated from the sampled values. If we can divide the set of sensors into disjoint so-called representative sampling subsets that each represent the other sensors sufficiently well, we can alternate the sampling between the sampling subsets and thus, increase battery life significantly. In this paper, we formulate the problem of finding representative sampling subsets as a graph problem on a so-called sensor graph with the sensors as nodes. Our proposed solution, SubGraphSample, consists of two phases. In Phase-I, we create edges in the sensor graph based on the similarities between the time series of sensor values, analyzing six different techniques based on proven time series similarity metrics. In Phase-II, we propose two new techniques and extend four existing ones to find the maximal number of representative sampling subsets. Finally, we propose AutoSubGraphSample which auto-selects the best technique for Phase-I and Phase-II for a given dataset. Our extensive experimental evaluation shows that our approach can yield significant battery life improvements within realistic error bounds.

cs.SI↗

arXiv2022

Detecting Stance in Tweets : A Signed Network based Approach

Identifying user stance related to a political event has several applications, like determination of individual stance, shaping of public opinion, identifying popularity of government measures and many others. The huge volume of political discussions on social media platforms, like, Twitter, provide opportunities in developing automated mechanisms to identify individual stance and subsequently, scale to a large volume of users. However, issues like short text and huge variance in the vocabulary of the tweets make such exercise enormously difficult. Existing stance detection algorithms require either event specific training data or annotated twitter handles and therefore, are difficult to adapt to new events. In this paper, we propose a sign network based framework that use external information sources, like news articles to create a signed network of relevant entities with respect to a news event and subsequently use the same to detect stance of any tweet towards the event. Validation on 5,000 tweets related to 10 events indicates that the proposed approach can ensure over 6.5% increase in average F1 score compared to the existing stance detection approaches.

cs.SI↗

arXiv2022

SigGAN : Adversarial Model for Learning Signed Relationships in Networks

Signed link prediction in graphs is an important problem that has applications in diverse domains. It is a binary classification problem that predicts whether an edge between a pair of nodes is positive or negative. Existing approaches for link prediction in unsigned networks cannot be directly applied for signed link prediction due to their inherent differences. Further, additional structural constraints, like, the structural balance property of the signed networks must be considered for signed link prediction. Recent signed link prediction approaches generate node representations using either generative models or discriminative models. Inspired by the recent success of Generative Adversarial Network (GAN) based models which comprises of a discriminator and generator in several applications, we propose a Generative Adversarial Network (GAN) based model for signed networks, SigGAN. It considers the requirements of signed networks, such as, integration of information from negative edges, high imbalance in number of positive and negative edges and structural balance theory. Comparing the performance with state of the art techniques on several real-world datasets validates the effectiveness of SigGAN.

cs.SI↗

arXiv2019

Predicting Tomorrow's Headline using Today's Twitter Deliberations

Predicting the popularity of news article is a challenging task. Existing literature mostly focused on article contents and polarity to predict popularity. However, existing research has not considered the users' preference towards a particular article. Understanding users' preference is an important aspect for predicting the popularity of news articles. Hence, we consider the social media data, from the Twitter platform, to address this research gap. In our proposed model, we have considered the users' involvement as well as the users' reaction towards an article to predict the popularity of the article. In short, we are predicting tomorrow's headline by probing today's Twitter discussion. We have considered 300 political news article from the New York Post, and our proposed approach has outperformed other baseline models.

cs.SI↗