Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Pushpak BhattacharyyaSuggested from co-authorship
- Suggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works48 from public data
- DialogueGCN: A Graph Convolutional Neural Network for Emotion Recognition in Conversation725
Through the graph network, DialogueGCN addresses context propagation issues present in the current RNN-based methods, and empirically show that this method alleviates such issues, while outperforming the current state of the art on a number of benchmark emotion classification datasets.
- COSMIC: COmmonSense knowledge for eMotion Identification in Conversations448
This paper proposes COSMIC, a new framework that incorporates different elements of commonsense such as mental states, events, and causal relations, and build upon them to learn interactions between interlocutors participating in a conversation.
- MIME: MIMicking Emotions for Empathetic Response Generation271
This work argues that empathetic responses often mimic the emotion of the user to a varying degree, depending on its positivity or negativity and content, and introduces stochasticity into the emotion mixture that yields emotionally more varied emPathetic responses than the previous work.
- Multi-task Learning for Multi-modal Emotion Recognition and Sentiment Analysis217
A deep multi-task learning framework that jointly performs sentiment and emotion analysis both as well as a context-level inter-modal attention framework for simultaneously predicting the sentiment and expressed emotions of an utterance is presented.
- Text-to-Audio Generation using Instruction-Tuned LLM and Latent Diffusion Model213
This work adopts an instruction-tuned LLM Flan-T5 as the text encoder for text-to-audio (TTA) generation and finds that the latent diffusion model (LDM)-based approach TANGO outperforms the state-of-the-art AudioLDM on most metrics and stays comparable on the rest on AudioCaps test set.
- Contextual Inter-modal Attention for Multi-modal Sentiment Analysis186
A recurrent neural network based multi-modal attention framework that leverages the contextual information for utterance-level sentiment prediction that applies attention on multi- modal multi-utterance representations and tries to learn the contributing features amongst them.
- Recognizing Emotion Cause in Conversations174
This work introduces a new task highly relevant for (explainable) emotion-aware artificial intelligence: recognizing emotion cause in conversations, provides a new highly challenging publicly available dialogue-level dataset for this task, and gives strong baseline results on this dataset.
- Mustango: Toward Controllable Text-to-Music Generation171
The quality of the music generated by Mustango is state-of-the-art, and the controllability through music-specific text prompts greatly outperforms other models such as MusicGen and AudioLDM2.
- Tango 2: Aligning Diffusion-based Text-to-Audio Generations through Direct Preference Optimization165
This work fine-tune the publicly available Tango text-to-audio model using diffusion-DPO (direct preference optimization) loss on the authors' preference dataset and shows that it leads to improved audio output over Tango and AudioLDM2, in terms of both automatic- and manual-evaluation metrics.
- Text-to-Audio Generation using Instruction Guided Latent Diffusion Model135
This work adopts an instruction-tuned LLM Flan-T5 as the text encoder for text-to-audio generation-a task where the goal is to generate an audio from its textual description and outperforms the state-of-the-art AudioLDM on most metrics and stays comparable on the rest on AudioCaps test set.
- 121
- A Multilayer Perceptron based Ensemble Technique for Fine-grained Financial Sentiment Analysis115
A novel method for combining deep learning and classical feature based models using a Multi-Layer Perceptron (MLP) network for financial sentiment analysis using a classical supervised model based on Support Vector Regression (SVR).
- Music Genre Recognition Using Deep Neural Networks and Transfer Learning95
This work proposes a novel approach for music genre recognition using an ensemble of convolutional long short term memory based neural networks (CNN LSTM) and a transfer learning model that outperforms them and achieves new state of the art results.
- 87
- KinGDOM: Knowledge-Guided DOMain Adaptation for Sentiment Analysis71
A new framework is introduced, KinGDOM, which utilizes the ConceptNet knowledge graph to enrich the semantics of a document by providing both domain-specific and domain-general background concepts, and conditioning a popular domain-adversarial baseline method with these learned concepts helps improve its performance over state-of-the-art approaches.
- IITP at SemEval-2017 Task 5: An Ensemble of Deep Learning and Feature Based Models for Financial Sentiment Analysis37
An ensemble based model which combines state of the art deep learning sentiment analysis algorithms like Convolution Neural Network and Long Short Term Memory along with feature based models to identify optimistic or pessimistic sentiments associated with companies and stocks in financial texts is proposed.
- 33
- Improving Zero-Shot Learning Baselines with Commonsense Knowledge32
This work takes advantage of explicit relations between nodes defined in ConceptNet, a commonsense knowledge graph, to generate commonsense embeddings of the class labels by using a graph convolution network-based autoencoder.
- 31
- CIDER: Commonsense Inference for Dialogue Explanation and Reasoning30
This work introduces CIDER – a manually curated dataset that contains dyadic dialogue explanations in the form of implicit and explicit knowledge triplets inferred using contextual commonsense inference that can be conducive to improving several downstream applications.
- Exemplars-Guided Empathetic Response Generation Controlled by the Elements of Human Communication29
Three elements of human communication—emotional presence, interpretation, and exploration—and sentiment are additionally introduced using synthetic labels to guide the generation towards empathy and it is empirically shown that these approaches yield significant improvements in empathetic response quality in terms of both automated and human-evaluated metrics.
- Are Language Models Puzzle Prodigies? Algorithmic Puzzles Unveil Serious Challenges in Multimodal Reasoning27
This paper presents a new dataset, AlgoPuzzleVQA, designed to challenge and evaluate the capabilities of multimodal language models in solving algorithmic puzzles that necessitate both visual understanding, language understanding, and complex algorithmic reasoning.
- Exploring the Role of Context in Utterance-level Emotion, Act and Intent Classification in Conversations: An Empirical Study26
This paper explores and quantify the role of context for different aspects of a dialogue, namely emotion, dialogue act, and intent identification, using state-of-the-art dialogue understanding methods as baselines and employs various perturbations to distort the context of a given utterance.
- Evaluating LLMs’ Mathematical and Coding Competency through Ontology-guided Interventions24
A significant performance drop is shown across all the models against the perturbed questions, suggesting that the current LLMs lack robust problem solving skills and structured reasoning abilities in many areas, as defined by the ontology.
- Flacuna: Unleashing the Problem Solving Power of Vicuna using FLAN Fine-Tuning24
Experimental findings strongly indicate that the enhanced problem-solving abilities of the model, FLACUNA, are obtained through fine-tuning VICUNA on the FLAN dataset, leading to significant improvements across numerous benchmark datasets in INSTRUCTEVAL.
- Multiview Contextual Commonsense Inference: A New Dataset and Task16
This work creates CICEROv2, a dataset consisting of 8,351 instances from 2,379 dialogues, containing multiple human-written answers for each contextual commonsense inference question, representing a type of explanation on cause, subsequent event, motivation, and emotional reaction.
- 16
- The FACTS Leaderboard: A Comprehensive Benchmark for Large Language Model Factuality15
An online leaderboard suite and associated set of benchmarks that comprehensively evaluates the ability of language models to generate factually accurate text across diverse scenarios, designed to provide a robust and balanced assessment of a model's overall factuality.
- AlgoPuzzleVQA: Diagnosing Multimodal Reasoning Challenges of Language Models with Algorithmic Multimodal Puzzles13
This paper presents a new dataset, A LGO P UZZLE VQA, designed to challenge and evaluate the capabilities of multimodal language models in solving algorithmic puzzles that necessitate both visual understanding, language understanding, and complex algorithmic reasoning, and reveals that multimodal language models exhibit limited performance in puzzle-solving tasks.
- STaCK: Sentence Ordering with Temporal Commonsense Knowledge12
STaCK is introduced — a framework based on graph neural networks and temporal commonsense knowledge to model global information and predict the relative order of sentences and results are reported on five different datasets.
- Persuasive dialogue understanding: The baselines and negative results12
This paper demonstrates the limitations of a Transformer-based approach coupled with Conditional Random Field (CRF) for the task of persuasive strategy recognition and leverages inter- and intra-speaker contextual semantic features, as well as label dependencies to improve the recognition.
- Not All Votes Count! Programs as Verifiers Improve Self-Consistency of Language Models for Math Reasoning10
Prove, a simple yet effective framework that leverages translated programs derived from natural language solutions as a verification mechanism to filter out potentially incorrect reasoning paths before aggregating final answers, is introduced.
- 8
- 7
- Deep Ensemble Model with the Fusion of Character, Word and Lexicon Level Information for Emotion and Sentiment Prediction7
A novel neural network based architecture which incorporates character, word and lexicon level information to predict the degree of intensity for sentiment and emotion attains the state-of-the-art performance.
- The Jumping Reasoning Curve? Tracking the Evolution of Reasoning Performance in GPT-[n] and o-[n] Models on Multimodal Puzzles6
The evolution of the GPT-[n] and o-[n] series models are tracked on challenging multimodal puzzles from PuzzleVQA and AlgoPuzzleVQA, which demand fine-grained visual perception and shows that o-[n] series significantly outperform the GPT-[n] series and show strong scalability in multimodal reasoning.
- 6
- 6
- MixEval-X: Any-to-Any Evaluations from Real-World Data Mixtures6
This work introduces MixEval-X, the first any-to-any, real-world benchmark designed to optimize and standardize evaluations across diverse input and output modalities, and proposes multi-modal benchmark mixture and adaptation-rectification pipelines to reconstruct real-world task distributions.
- ReTAG: Reasoning Aware Table to Analytic Text Generation6
ReTAG is the first model that can controllably use multiple reasoning methods within a structure-aware sequence to sequence model to surpass state of the art performance in multiple table to text tasks.
- 4
- Discovering the complete enhancer map of human herpesviruses using a natural language processing model1
An AI tool is presented that successfully predicts enhancers in both viral and human genomes and can identify known enhancers and predict novel enhancer elements in other herpesviruses, different viruses, and the human genome.
- Generating Intermediate Steps for NLI with Next-Step Supervision1
This work trains a sequence-to-sequence model to generate only the next step given an NLI premise and hypothesis pair; enhances it with external knowledge and symbolic search to generate intermediate steps with only next-step supervision and shows the correctness of such generated steps through automated and human verification.
- Test-Time Scaling via Error Localization–
This work introduces Test-Time Scaling via Error Localization (TTEL), an inference-time algorithm that utilizes fixed or environment feedback to perform token-level error localization and establishes strictly dominating Pareto frontiers across sequential reasoning domains.
- –
- Prover: Generating Intermediate Steps for NLI with Commonsense Knowledge Retrieval and Next-Step Prediction–
This work trains and enhances a sequence-to-sequence next-step prediction model with external commonsense knowledge and search to generate intermediate steps with limited next-step supervision and shows that such generated steps can help improve end-to-end NLI task performance using simple data augmentation strategies.
- Vector Space Interpolation for Query Expansion–
This work describes a method for query set expansion about emerging topics using vector space interpolation using a transformer model called OPTIMUS, which is suitable for vector space manipulation due to its variational autoencoder nature.
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-12. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.