Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.15050 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917283446652928 |
|---|---|
| author | Kumar, Gaurav Garg, Ayush Mazumder, Debajyoti Kishore, Aditya kumar, Babu Patro, Jasabanta |
| author_facet | Kumar, Gaurav Garg, Ayush Mazumder, Debajyoti Kishore, Aditya kumar, Babu Patro, Jasabanta |
| contents | Automated fact-checking has been a challenging task for the research community. Prior work has explored various strategies, such as end-to-end training, retrieval-augmented generation, and prompt engineering, to build robust fact-checking systems. However, their accuracy has not been high enough for real-world deployment. We, on the other hand, propose a new learning paradigm, where evidence classification and entailed justifications made by generative language models (GLMs) are used to train encoder-only language models (ELMs). We conducted a rigorous set of experiments, comparing our approach with recent works along with various prompting and fine-tuning strategies. Additionally, we performed ablation studies, error analysis, quality analysis of model explanations, and a domain generalisation study to provide a comprehensive understanding of our approach. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_15050 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Entailed Opinion Matters: Improving the Fact-Checking Performance of Language Models by Relying on their Entailment Ability Kumar, Gaurav Garg, Ayush Mazumder, Debajyoti Kishore, Aditya kumar, Babu Patro, Jasabanta Computation and Language Automated fact-checking has been a challenging task for the research community. Prior work has explored various strategies, such as end-to-end training, retrieval-augmented generation, and prompt engineering, to build robust fact-checking systems. However, their accuracy has not been high enough for real-world deployment. We, on the other hand, propose a new learning paradigm, where evidence classification and entailed justifications made by generative language models (GLMs) are used to train encoder-only language models (ELMs). We conducted a rigorous set of experiments, comparing our approach with recent works along with various prompting and fine-tuning strategies. Additionally, we performed ablation studies, error analysis, quality analysis of model explanations, and a domain generalisation study to provide a comprehensive understanding of our approach. |
| title | Entailed Opinion Matters: Improving the Fact-Checking Performance of Language Models by Relying on their Entailment Ability |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2505.15050 |