Distill and Align Decomposition for Enhanced Claim Verification
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908852237107200 |
|---|---|
| author | Magomere, Jabez Kochkina, Elena Mensah, Samuel Kaur, Simerjot Acero, Fernando Oncevay, Arturo Smiley, Charese H. Liu, Xiaomo Veloso, Manuela |
| author_facet | Magomere, Jabez Kochkina, Elena Mensah, Samuel Kaur, Simerjot Acero, Fernando Oncevay, Arturo Smiley, Charese H. Liu, Xiaomo Veloso, Manuela |
| contents | Complex claim verification requires decomposing sentences into verifiable subclaims, yet existing methods struggle to align decomposition quality with verification performance. We propose a reinforcement learning (RL) approach that jointly optimizes decomposition quality and verifier alignment using Group Relative Policy Optimization (GRPO). Our method integrates: (i) structured sequential reasoning; (ii) supervised finetuning on teacher-distilled exemplars; and (iii) a multi-objective reward balancing format compliance, verifier alignment, and decomposition quality. Across six evaluation settings, our trained 8B decomposer improves downstream verification performance to (71.75%) macro-F1, outperforming prompt-based approaches ((+1.99), (+6.24)) and existing RL methods ((+5.84)). Human evaluation confirms the high quality of the generated subclaims. Our framework enables smaller language models to achieve state-of-the-art claim verification by jointly optimising for verification accuracy and decomposition quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_21857 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Distill and Align Decomposition for Enhanced Claim Verification Magomere, Jabez Kochkina, Elena Mensah, Samuel Kaur, Simerjot Acero, Fernando Oncevay, Arturo Smiley, Charese H. Liu, Xiaomo Veloso, Manuela Artificial Intelligence Computation and Language Machine Learning Complex claim verification requires decomposing sentences into verifiable subclaims, yet existing methods struggle to align decomposition quality with verification performance. We propose a reinforcement learning (RL) approach that jointly optimizes decomposition quality and verifier alignment using Group Relative Policy Optimization (GRPO). Our method integrates: (i) structured sequential reasoning; (ii) supervised finetuning on teacher-distilled exemplars; and (iii) a multi-objective reward balancing format compliance, verifier alignment, and decomposition quality. Across six evaluation settings, our trained 8B decomposer improves downstream verification performance to (71.75%) macro-F1, outperforming prompt-based approaches ((+1.99), (+6.24)) and existing RL methods ((+5.84)). Human evaluation confirms the high quality of the generated subclaims. Our framework enables smaller language models to achieve state-of-the-art claim verification by jointly optimising for verification accuracy and decomposition quality. |
| title | Distill and Align Decomposition for Enhanced Claim Verification |
| topic | Artificial Intelligence Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2602.21857 |