Diffusion of Thoughts: Chain-of-Thought Reasoning in Diffusion Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915048527495168 |
|---|---|
| author | Ye, Jiacheng Gong, Shansan Chen, Liheng Zheng, Lin Gao, Jiahui Shi, Han Wu, Chuan Jiang, Xin Li, Zhenguo Bi, Wei Kong, Lingpeng |
| author_facet | Ye, Jiacheng Gong, Shansan Chen, Liheng Zheng, Lin Gao, Jiahui Shi, Han Wu, Chuan Jiang, Xin Li, Zhenguo Bi, Wei Kong, Lingpeng |
| contents | Recently, diffusion models have garnered significant interest in the field of text processing due to their many potential advantages compared to conventional autoregressive models. In this work, we propose Diffusion-of-Thought (DoT), a novel approach that integrates diffusion models with Chain-of-Thought, a well-established technique for improving the reasoning ability of autoregressive language models. In contrast to autoregressive language models that make decisions in a left-to-right, token-by-token manner, DoT allows reasoning steps to diffuse over time through a diffusion language model and offers greater flexibility in trading-off computation for reasoning performance. Our experimental results demonstrate the effectiveness of DoT in multi-digit multiplication, boolean logic, and grade school math problems, with a small diffusion model outperforming a much larger autoregressive model in both efficiency and accuracy. In addition to that, DoT showcases promising self-correction abilities and benefits from existing reasoning-enhancing techniques like self-consistency decoding. Our findings contribute to the understanding and development of reasoning with diffusion language models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_07754 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Diffusion of Thoughts: Chain-of-Thought Reasoning in Diffusion Language Models Ye, Jiacheng Gong, Shansan Chen, Liheng Zheng, Lin Gao, Jiahui Shi, Han Wu, Chuan Jiang, Xin Li, Zhenguo Bi, Wei Kong, Lingpeng Computation and Language Artificial Intelligence Machine Learning Recently, diffusion models have garnered significant interest in the field of text processing due to their many potential advantages compared to conventional autoregressive models. In this work, we propose Diffusion-of-Thought (DoT), a novel approach that integrates diffusion models with Chain-of-Thought, a well-established technique for improving the reasoning ability of autoregressive language models. In contrast to autoregressive language models that make decisions in a left-to-right, token-by-token manner, DoT allows reasoning steps to diffuse over time through a diffusion language model and offers greater flexibility in trading-off computation for reasoning performance. Our experimental results demonstrate the effectiveness of DoT in multi-digit multiplication, boolean logic, and grade school math problems, with a small diffusion model outperforming a much larger autoregressive model in both efficiency and accuracy. In addition to that, DoT showcases promising self-correction abilities and benefits from existing reasoning-enhancing techniques like self-consistency decoding. Our findings contribute to the understanding and development of reasoning with diffusion language models. |
| title | Diffusion of Thoughts: Chain-of-Thought Reasoning in Diffusion Language Models |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2402.07754 |