Beyond Mode Collapse: Distribution Matching for Diverse Reasoning
The paper discusses a new approach to on-policy reinforcement learning that addresses the issue of mode collapse. The proposed method, DMPO, enhances solution diversity by aligning policy distributions with a target distribution based on rewards. This approach has shown significant improvements in various reasoning tasks, demonstrating its effectiveness in maintaining exploration during training.
- ▪On-policy reinforcement learning methods like GRPO often suffer from mode collapse, leading to reduced solution diversity.
- ▪The proposed DMPO method prevents mode collapse by approximating forward KL minimization and aligning policy distributions with a target distribution based on rewards.
- ▪DMPO achieved notable improvements in quality ratios on NP-hard combinatorial optimization tasks, outperforming GRPO in both text-based and vision-based benchmarks.
arXiv cs.AI files mainly under ai research. We currently carry 1,128 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | arXiv cs.AI |
| Canonical URL | https://arxiv.org/abs/2605.19461 |
| Publication time | Wed, 20 May 2026 00:00:00 -0400 |
| Retrieval time | 2026-05-20T04:04:59.484Z |
| Last seen | 2026-05-20T04:04:59.484Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | mjbsF2bheKkY |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.19461 (cs) [Submitted on 19 May 2026] Title:Beyond Mode Collapse: Distribution Matching for Diverse Reasoning Authors:Xiaozhe Li, Yang Li, Xinyu Fang, Shengyuan Ding, Peiji Li, Yongkang Chen, Yichuan Ma, Tianyi Lyu, Linyang Li, Dahua Lin, Qipeng Guo, Qingwen Liu, Kai Chen View a PDF of the paper titled Beyond Mode Collapse: Distribution Matching for Diverse Reasoning, by Xiaozhe Li and 12 other authors View PDF HTML (experimental) Abstract:On-policy reinforcement learning methods like GRPO suffer from mode collapse: they exhibit reduced solution diversity, concentrating probability mass on a single solution once discovered and ceasing exploration of alternative strategies.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
