Why Sample What You Can Enumerate? Exact Policy Optimization for Genomic Tool Selection
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
Public signals
Semantic Scholar citations 0 · Semantic Scholar influential citations 0
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.