Papers by Debanjan Ghosh
The Benefits of Label-Description Training for Zero-Shot Text Classification (2023.emnlp-main)
Copied to clipboard
| Challenge: | Pretrained language models have improved zero-shot text classification by allowing the transfer of semantic knowledge from the training data to classify among specific label sets in downstream tasks. |
| Approach: | They propose to use a small finetuning dataset to describe the labels for a task and to use it to further improve zero-shot accuracies. |
| Outcome: | The proposed model is more accurate than zero-shot by 17-19% absolute across topic and sentiment datasets and more robust to choices required for zero- shot classification. |
Investigating the Impact of Conceptual Metaphors on LLM-based NLI through Shapley Interactions (2025.findings-emnlp)
Copied to clipboard
Meghdut Sengupta, Maximilian Muschalik, Fabian Fumagalli, Barbara Hammer, Eyke Hüllermeier, Debanjan Ghosh, Henning Wachsmuth
| Challenge: | Metaphorical language is prevalent in everyday communication, often used unconsciously, as in "rising crime" incorporating conceptual metaphors improves performance on downstream tasks that implicitly require correct metaphor interpretation, such as NLI. |
| Approach: | They extend metaphorical texts in an existing NLI dataset by source and target domains and conduct an ablation study using Shapley values and interactions to assess the extent to which LLMs interpret metaphorical language correctly in NLI. |
| Outcome: | The proposed method improves performance on natural language inference tasks with metaphoric input. |
AGReE: A system for generating Automated Grammar Reading Exercises (2022.emnlp-demos)
Copied to clipboard
| Challenge: | AGReE is a system that generates multiple-choice grammar practice items . common core standards for K-12 English literacy include grammar as a learning outcome . |
| Approach: | They propose a system that generates multiple-choice grammar practice exercises that can be completed while reading. |
| Outcome: | The proposed grammar-reading exercise system can be completed while reading . it offers immediate feedback, similar to a more formal incentive system . |
Figurative Language in Recognizing Textual Entailment (2021.findings-acl)
Copied to clipboard
| Challenge: | Existing RTE models struggle to capture figurative language, despite its ubiquity, it remains a bottleneck in automatic text understanding. |
| Approach: | They propose to frame five existing figurative language datasets into over 12,500 RTE examples. |
| Outcome: | The proposed models struggle to perform pragmatic inference and reasoning about world knowledge. |
“What makes a question inquisitive?” A Study on Type-Controlled Inquisitive Question Generation (2022.starsem-1)
Copied to clipboard
| Challenge: | Empirical results demonstrate that we can generate a variety of questions that adhere to specific types while drawing from the source texts. |
| Approach: | They propose a type-controlled framework for inquisitive question generation . they annotate an inquisite question dataset and train question type classifiers . |
| Outcome: | The proposed framework generates questions that adhere to specific types while drawing from the source texts. |
Rˆ3: Reverse, Retrieve, and Rank for Sarcasm Generation with Commonsense Knowledge (2020.acl-main)
Copied to clipboard
| Challenge: | Existing work on sarcasm generation focuses on context incongruity, but new work addresses this problem . |
| Approach: | They propose an unsupervised approach for sarcasm generation based on a non-sarcastic input sentence. |
| Outcome: | The proposed method generates sarcasm better than humans 34% of the time and better than a reinforced hybrid baseline 90% of the times. |
“Laughing at you or with you”: The Role of Sarcasm in Shaping the Disagreement Space (2021.eacl-main)
Copied to clipboard
| Challenge: | Detecting arguments in online interactions is useful to understand how conflicts arise and get resolved. |
| Approach: | They propose to use a corpus annotated with argumentative moves and sarcasm to model sarcastic relationships using deep learning architectures. |
| Outcome: | The proposed setup improves the argumentative relation classification task using deep learning architectures. |
FLUTE: Figurative Language Understanding through Textual Explanations (2022.emnlp-main)
Copied to clipboard
| Challenge: | Figurative language understanding is a recognizing textual entailment task, but lacks data for figurative language. |
| Approach: | They propose to use a dataset to analyze figurative NLI instances with explanations to improve models' performance. |
| Outcome: | The proposed dataset can scale up models even for figurative language using human annotations. |
Controlled Language Generation for Language Learning Items (2022.emnlp-industry)
Copied to clipboard
| Challenge: | Recent advances in pre-trained language models have resulted in success in generating fluent English text. |
| Approach: | They propose to employ natural language generation to rapidly generate English language items . they experiment with deep pretrained models and develop methods for controlling items for factors relevant in language learning . |
| Outcome: | The proposed framework shows high grammatically scores for all models and higher complexity over baseline models. |