Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2310.18794 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916204050907136 |
|---|---|
| author | Wan, Yixin Wu, Fanyou Xu, Weijie Sengamedu, Srinivasan H. |
| author_facet | Wan, Yixin Wu, Fanyou Xu, Weijie Sengamedu, Srinivasan H. |
| contents | In this work, we propose sequence-level certainty as a common theme over hallucination in Knowledge Grounded Dialogue Generation (KGDG). We explore the correlation between the level of hallucination in model responses and two types of sequence-level certainty: probabilistic certainty and semantic certainty. Empirical results reveal that higher levels of both types of certainty in model responses are correlated with lower levels of hallucination. We further propose Certainty-based Response Ranking (CRR), a decoding-time hallucination mitigation method that samples several response candidates, ranks them based on sequence-level certainty, and outputs the response with the highest certainty level. Aligning with our definitions of sequence-level certainty, we design 2 types of CRR approaches: Probabilistic CRR (P-CRR) and Semantic CRR (S-CRR). P-CRR ranks individually sampled model responses using the arithmetic mean log-probability of the entire sequence. S-CRR approaches certainty estimation from meaning-space, and ranks model response candidates based on their semantic certainty level as measured by an entailment-based Agreement Score (AS). Through extensive experiments across 3 KGDG datasets, 3 decoding methods, and 4 KGDG models, we validate the effectiveness of CRR for reducing hallucination in KGDG task. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2310_18794 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Sequence-Level Certainty Reduces Hallucination In Knowledge-Grounded Dialogue Generation Wan, Yixin Wu, Fanyou Xu, Weijie Sengamedu, Srinivasan H. Computation and Language Artificial Intelligence In this work, we propose sequence-level certainty as a common theme over hallucination in Knowledge Grounded Dialogue Generation (KGDG). We explore the correlation between the level of hallucination in model responses and two types of sequence-level certainty: probabilistic certainty and semantic certainty. Empirical results reveal that higher levels of both types of certainty in model responses are correlated with lower levels of hallucination. We further propose Certainty-based Response Ranking (CRR), a decoding-time hallucination mitigation method that samples several response candidates, ranks them based on sequence-level certainty, and outputs the response with the highest certainty level. Aligning with our definitions of sequence-level certainty, we design 2 types of CRR approaches: Probabilistic CRR (P-CRR) and Semantic CRR (S-CRR). P-CRR ranks individually sampled model responses using the arithmetic mean log-probability of the entire sequence. S-CRR approaches certainty estimation from meaning-space, and ranks model response candidates based on their semantic certainty level as measured by an entailment-based Agreement Score (AS). Through extensive experiments across 3 KGDG datasets, 3 decoding methods, and 4 KGDG models, we validate the effectiveness of CRR for reducing hallucination in KGDG task. |
| title | Sequence-Level Certainty Reduces Hallucination In Knowledge-Grounded Dialogue Generation |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2310.18794 |