{"ID":2862043,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.00805","arxiv_id":"2510.00805","title":"Planning-Augmented Sampling with Early Guidance for High-Reward Discovery","abstract":"Generative Flow Networks (GFlowNets) enable structured generation with inherent diversity, but existing sampling strategies often rely on weak guided exploration, slowing early discovery of high-reward candidates. In tasks such as molecular design, rapid and consistent generation of high-reward solutions can outweigh faithful distribution matching. We propose a planning-augmented framework in which Monte Carlo Tree Search using polynomial upper confidence bounds provides online value estimates, and a controllable soft-greedy mechanism integrates these planning signals into the GFlowNets forward policy. This design fosters early exploration of high-reward trajectories and gradually shifts to policy-driven exploitation as experience accumulates. Empirical results show that our method accelerates early high-reward discovery, sustains top-quality sample generation, and preserves diversity across representative tasks. All implementations are available at https://github.com/ZRNB/PLUS.","short_abstract":"Generative Flow Networks (GFlowNets) enable structured generation with inherent diversity, but existing sampling strategies often rely on weak guided exploration, slowing early discovery of high-reward candidates. In tasks such as molecular design, rapid and consistent generation of high-reward solutions can outweigh f...","url_abs":"https://arxiv.org/abs/2510.00805","url_pdf":"https://arxiv.org/pdf/2510.00805v3","authors":"[\"Rui Zhu\",\"Yudong Zhang\",\"Xuan Yu\",\"Chen Zhang\",\"Xu Wang\",\"Yang Wang\"]","published":"2025-10-01T12:09:04Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\"]","methods":"[\"LoRA\"]","has_code":false,"code_links":[{"ID":608867,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2862043,"paper_url":"https://arxiv.org/abs/2510.00805","paper_title":"Planning-Augmented Sampling with Early Guidance for High-Reward Discovery","repo_url":"https://github.com/ZRNB/PLUS","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}