Papers by Deepanway Ghosal

19 papers
NLKI: A Lightweight Natural Language Knowledge Integration Framework for Improving Small VLMs in Commonsense VQA Tasks (2025.findings-emnlp)

Copied to clipboard

Challenge: Small vision-language models lag behind their larger generative counterparts due to lack of knowledge.
Approach: They propose a framework that integrates commonsense knowledge into small vision-language models . the framework retrieves natural language facts and prompts an LLM to craft natural language explanations .
Outcome: The proposed framework retrieves natural language facts and prompts an LLM to craft natural language explanations.
PuzzleVQA: Diagnosing Multimodal Reasoning Challenges of Language Models with Abstract Visual Patterns (2024.findings-acl)

Copied to clipboard

Challenge: recognizing patterns and abstracting concepts are key to general intelligence, we show . state-of-the-art large multimodal models struggle to generalize well to simple abstract patterns .
Approach: They evaluate large multimodal models with abstract patterns based on colors, numbers, sizes, and shapes.
Outcome: The proposed model fails to generalize well to simple abstract patterns, the study shows . the model fails on single-concept puzzles, despite its sophistication .
COSMIC: COmmonSense knowledge for eMotion Identification in Conversations (2020.findings-emnlp)

Copied to clipboard

Challenge: Current methods for emotion recognition in conversations often face difficulties in context propagation, emotion shift detection, and differentiating between related emotion classes.
Approach: They propose a framework that incorporates mental states, events, and causal relations to learn interactions between interlocutors participating in a conversation.
Outcome: The proposed framework improves on four conversational benchmark datasets.
Exploring the Role of Context in Utterance-level Emotion, Act and Intent Classification in Conversations: An Empirical Study (2021.findings-acl)

Copied to clipboard

Challenge: utterance-level dialogue understanding tasks are often performed at utterrance level and are often conjoined together under the umbrella of utterence-level dialog understanding.
Approach: They propose to use a contextual utterance-level dialogue understanding baseline as a strong framework for six dialogue-understanding tasks.
Outcome: The proposed framework can be easily adapted for other tasks for similar purposes.
Language Guided Visual Question Answering: Elevate Your Multimodal Language Model Using Knowledge-Enriched Prompts (2023.findings-emnlp)

Copied to clipboard

Challenge: Visual question answering (VQA) is a task that requires an understanding of both the image and the question to provide a natural language answer.
Approach: They propose a multimodal framework that leverages language guidance to answer questions more accurately.
Outcome: The proposed framework improves on the multi-choice question-answering task using CLIP and BLIP models.
MIME: MIMicking Emotions for Empathetic Response Generation (2020.emnlp-main)

Copied to clipboard

Challenge: Empathy is a fundamental human trait that reflects our ability to understand and reflect the thoughts and feelings of the people we interact with.
Approach: They propose to use polarity-based emotion clusters to generate empathetic responses . they also introduce stochasticity into the emotion mixture that yields emotionally more varied responses compared to the previous work .
Outcome: The proposed methods improve empathy and contextual relevance of the response, and introduce stochasticity into the emotion mixture that yields emotionally more varied responses than the previous work.
Vector Space Interpolation for Query Expansion (2022.aacl-short)

Copied to clipboard

Challenge: Topic-sensitive query set expansion is crucial for queries related to sensitive and emerging topics.
Approach: They propose a method for topic-sensitive query set expansion using vector space interpolation.
Outcome: The proposed method generates new queries about the sensitive topic by incorporating set diversity, which is not captured by traditional sentence-level augmentation methods such as paraphrasing or back-translation.
Contextual Inter-modal Attention for Multi-modal Sentiment Analysis (D18-1)

Copied to clipboard

Challenge: Existing methods for multi-modal sentiment analysis are limited due to the use of text, visual and acoustic inputs.
Approach: They propose a recurrent neural network based multi-modal attention framework that leverages contextual information for utterance-level sentiment prediction.
Outcome: The proposed framework performs better on two multi-modal sentiment analysis benchmark datasets with accuracies of 82.31% and 79.80% for the MOSI and MOSEI datasets.
ReTAG: Reasoning Aware Table to Analytic Text Generation (2023.emnlp-main)

Copied to clipboard

Challenge: Table to text models generate descriptive summaries that repeat information contained within a table in sentences.
Approach: They propose a table-aware table-to-text model that uses vector-quantization to infuse different types of analytical reasoning into the output.
Outcome: The proposed model achieves 2.2%, 2.9% improvement on PARENT metric over state-of-the-art models.
CICERO: A Dataset for Contextualized Commonsense Inference in Dialogues (2022.acl-long)

Copied to clipboard

Challenge: Fig. 1a shows an example where commonsense knowledge is crucial in sifting relevant information from the context.
Approach: They curate a dataset of dyadic conversations with five types of utterance-level reasoning-based inferences: cause, subsequent event, prerequisite, motivation, and emotional reaction.
Outcome: The dataset contains 53,105 of such inferences from 5,672 dialogues.
Emma-X: An Embodied Multimodal Action Model with Grounded Chain of Thought and Look-ahead Spatial Reasoning (2025.acl-long)

Copied to clipboard

Challenge: Visual-Language-Action models lack the ability to generate actionable policies tailored to specific robotic embodiments.
Approach: They propose an embodied multimodal action model with Grounded Chain of Thought and Look-ahead Spatial Reasoning that enhances spatial reasoning and task planning.
Outcome: The proposed model improves on existing baselines in tasks requiring spatial reasoning and grounding reasoning.
STaCK: Sentence Ordering with Temporal Commonsense Knowledge (2021.emnlp-main)

Copied to clipboard

Challenge: Existing approaches to sentence order prediction ignore the importance of document level global information, i.e., while predicting relative order of two sentences (s i , s j) other sentences sk from the same document do not play any role.
Approach: They propose a framework based on graph neural networks and temporal commonsense knowledge to model global information and predict relative order of sentences.
Outcome: The proposed method is naturally suitable for order prediction on five different datasets and has potential applications in the evaluation of the quality of machinegenerated documents.
Evaluating LLMs’ Mathematical and Coding Competency through Ontology-guided Interventions (2025.findings-acl)

Copied to clipboard

Challenge: Current large language models have shown impressive performance on logical reasoning benchmarks . however, the true depth of their competencies and robustness in reasoning tasks remains an open question .
Approach: They propose a general ontology of perturbations and a semi-automatic method to apply perturbations to arithmetic reasoning and code generation datasets to test their LLMs' capabilities.
Outcome: The proposed model outperforms existing models on arithmetic reasoning and code generation tasks.
DialogueGCN: A Graph Convolutional Neural Network for Emotion Recognition in Conversation (D19-1)

Copied to clipboard

Challenge: Emotion recognition in conversation (ERC) has received much attention lately due to its potential widespread applications in diverse areas, such as health-care, education, and human resources.
Approach: They propose a graph neural network-based approach to emotion recognition in conversation that leverages self and inter-speaker dependency of the interlocutors to model conversational context.
Outcome: The proposed method outperforms the current state-of-the-art on a number of benchmark emotion classification datasets while minimizing context propagation issues.
AlgoPuzzleVQA: Diagnosing Multimodal Reasoning Challenges of Language Models with Algorithmic Multimodal Puzzles (2025.naacl-long)

Copied to clipboard

Challenge: Existing datasets focused on visual question-answering focus on visual, language, and algorithmic knowledge . a new study examines the performance of multimodal language models in solving algorithmic puzzles .
Approach: They propose a dataset to test the capabilities of multimodal language models in solving algorithmic puzzles.
Outcome: The proposed dataset is generated automatically from human code.
Mustango: Toward Controllable Text-to-Music Generation (2024.naacl-long)

Copied to clipboard

Challenge: Mustango is a text-to-music system that allows music-domain-knowledge-informed text-based music generation.
Approach: They propose a music-domain-knowledge-inspired text-to-music system based on diffusion that generates music with captions that include specific instructions related to chords, beats, key and tempo.
Outcome: The proposed system outperforms existing models in music generation tasks.
KinGDOM: Knowledge-Guided DOMain Adaptation for Sentiment Analysis (2020.acl-main)

Copied to clipboard

Challenge: Existing approaches to cross-domain sentiment analysis cannot be reliably deployed due to the distributional mismatch between training and evaluation domains.
Approach: They propose a framework that uses ConceptNet to enrich semantics of documents by providing domain-specific and domain-general background concepts.
Outcome: The proposed framework improves on a domain-adversarial baseline method and can be used in domain adaptation.
Two is Better than Many? Binary Classification as an Effective Approach to Multi-Choice Question Answering (2022.emnlp-main)

Copied to clipboard

Challenge: Existing approaches to multi-choice question answering are based on binary classifications instead of scoring each answer as a single class.
Approach: They propose a simple refactoring of multi-choice question answering tasks as a series of binary classifications and propose re-framing to make them more efficient.
Outcome: The proposed approach is significantly more effective across different tasks and models.
Multi-task Learning for Multi-modal Emotion Recognition and Sentiment Analysis (N19-1)

Copied to clipboard

Challenge: Existing frameworks for sentiment and emotion analysis are not efficient for inter-task learning.
Approach: They propose a multi-task learning framework that performs sentiment and emotion analysis together.
Outcome: The proposed framework improves on a CMU-MOSEI dataset for sentiment and emotion analysis.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations