Papers by Vikram Goyal
Tox-BART: Leveraging Toxicity Attributes for Explanation Generation of Implicit Hate Speech (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing language models to generate implicit hate explanations are lacking in many fields. |
| Approach: | They propose to use language models to generate explicit hate posts to make it clear . they find that simpler models incorporating external toxicity signals outperform KG-infused models . |
| Outcome: | The proposed setup produces more precise explanations than zero-shot GPT-3.5, highlighting the intricate nature of the task. |
Probing Critical Learning Dynamics of PLMs for Hate Speech Detection (2024.findings-eacl)
Copied to clipboard
| Challenge: | Existing studies on pretrained language models (PLMs) for hate speech detection have not investigated how their performance is affected by pretraining and finetuning. |
| Approach: | They propose to compare pretrained language models, evaluate their seed robustness, finetuning settings, and the impact of pretraining data collection time. |
| Outcome: | The proposed models show that they are more robust than other models and that they have a better chance of performing better than domain-specific models. |
SymTax: Symbiotic Relationship and Taxonomy Fusion for Effective Citation Recommendation (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing recommendations focus on local context or global context but fail to consider actual human citation behaviour. |
| Approach: | They propose a recommendation architecture that considers both local and global contexts . they use hyperbolic separation to compute query-candidate similarity . |
| Outcome: | The proposed framework performs better on a large dataset with 8.27 million citation contexts . it learns to embed the infused taxonomies in the hyperbolic space and computes similarity . |
Multi-Relational Graph Transformer for Automatic Short Answer Grading (2022.naacl-main)
Copied to clipboard
| Challenge: | Existing methods for Automatic Short Answer Grading (ASAG) ignore structural context and therefore do not perform well. |
| Approach: | They propose a Multi-Relational Graph Transformer to prepare token representations considering the structural context of a sentence. |
| Outcome: | The proposed model outperforms existing state-of-the-art methods on a dataset from an undergraduate computer science course. |
Promoting Topic Coherence and Inter-Document Consorts in Multi-Document Summarization via Simplicial Complex and Sheaf Graph (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing systems that generate summaries from multiple sources often lack accuracy and accuracy due to the length of tokens used in encoding. |
| Approach: | They propose a novel encoder-decoder model that uses pre-trained BART to analyze linguistic nuances, simplicial complex layer to apprehend inherent properties that transcend pairwise associations and sheaf graph attention to effectively capture heterophilic properties. |
| Outcome: | The proposed model achieves consistent performance improvement across all evaluation metrics (syntactical, semantical and faithfulness). |