🛰️ Daily AI Frontier
‹ back to 2026-09-10

Why Sample What You Can Enumerate? Exact Policy Optimization for Genomic Tool Selection

Research Bioinformatics AI

Ranking

Overall 78
Content 95
Popularity 37

Observed public metrics from 1 member.

Merged summary

TL;DR - FGPO exactly enumerates and scores all tool subsets when training frozen genomic reasoners, avoiding GRPO’s increasingly weak reward signal. It outperformed GRPO across 15 benchmark settings while reducing reward-evaluation costs and unnecessary tool use.

  • GRPO’s sampled, group-normalized advantages can vanish as the policy repeatedly selects preferred subsets; genomic questions with no reward signal rose from 0.2% to 20.8% after training.
  • FGPO optimizes the exact expectation over the complete action space and precomputes rewards for every question–subset pair.
  • Across five frozen reasoners and three genomic benchmarks, FGPO beat GRPO in all 15 settings by 6.75 points on average and up to 14.20 points.
  • Compared with on-demand GRPO, FGPO avoided a schedule requiring 2.4Ă— more reward evaluations and reduced GenomeQA tool calls from 2.36 to 1.40 per question.

Sources (1)

Why Sample What You Can Enumerate? Exact Policy Optimization for Genomic Tool Selection

arXiv cs.AI Haoyue Liu, Xiaoyu Ma, Ye Chen, Zhichao Wang, Xiaoying Tang 2026-09-09 arXiv:2609.10221
Public signals Semantic Scholar citations 0 · Semantic Scholar influential citations 0
Providers: Hugging Face · N/A OpenAlex · N/A Publisher · N/A Semantic Scholar · Citations 0 · Influential citations 0 X · N/A Fetched 2026-09-19 14:15:29.669334 UTC

TL;DR - FGPO exactly enumerates and scores all tool subsets when training frozen genomic reasoners, avoiding GRPO’s increasingly weak reward signal. It outperformed GRPO across 15 benchmark settings while reducing reward-evaluation costs and unnecessary tool use.

  • GRPO’s sampled, group-normalized advantages can vanish as the policy repeatedly selects preferred subsets; genomic questions with no reward signal rose from 0.2% to 20.8% after training.
  • FGPO optimizes the exact expectation over the complete action space and precomputes rewards for every question–subset pair.
  • Across five frozen reasoners and three genomic benchmarks, FGPO beat GRPO in all 15 settings by 6.75 points on average and up to 14.20 points.
  • Compared with on-demand GRPO, FGPO avoided a schedule requiring 2.4Ă— more reward evaluations and reduced GenomeQA tool calls from 2.36 to 1.40 per question.
item →