AI Paper Atlas

Research map / Large language models / Agents, code and math

BOOSTAPR: Boosting Automated Program Repair via Execution-Grounded Reinforcement Learning with Dual Reward Models

Yuanhao Li, Hongbo Wang, Xiaotang Shang, Xunzhu Tang, Yiming Cao, Xuhong Chen

ICML 2026 · Large language models › Agents, code and math · approach: LLM-guided program synthesis · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add PPO with sequence-level reward model Rseq after SFT (ablation)BoostAPR (Qwen2.5-Coder-32B) · SWE-bench Verifiedpass@1 (%): 23.4 -> 38.3Improved
add execution-verified SFT with reasoning traces (ablation)Qwen2.5-Coder-32B base · SWE-bench Verifiedpass@1 (%): 17.8 -> 23.4Improved
add line-level credit allocator Rline to PPO with Rseq (ablation)BoostAPR · SWE-bench Verifiedpass@1 (%): 38.3 -> 40.7Improved
add line-level credit allocator Rline to PPO with Rseq (ablation)BoostAPR · Defects4J v2.0 (Java transfer)resolve rate (%): 19.2 -> 24.8Improved
add line-level credit allocator Rline to PPO with Rseq (ablation)BoostAPR · QuixBugsfixed bugs (%): 95.0 -> 95.0No clear effect
use Rseq only for reranking instead of PPO training (ablation)SFT policy · SWE-bench Verifiedpass@1 (%): 26.1 vs 38.3 with PPOHurt
use edit-line spans instead of token-uniform credit (ablation)BoostAPR · SWE-bench Verifiedresolve rate (%): 37.6 -> 40.7Improved
use edit-line spans instead of hunk-level credit (ablation)BoostAPR · Defects4J v2.0 (Java transfer)resolve rate (%): 21.3 -> 24.8Improved
use edit-line spans instead of statement-level credit (ablation)BoostAPR · Defects4J v2.0 (Java transfer)resolve rate (%): 22.1 -> 24.8Improved
score patches only (no bug context) in Rseq (ablation)Rseq reward model · SWE-bench Verifiedpass@1 (%): 38.9 -> 40.7Improved
train Rseq with hybrid regression plus preference objective (ablation)Rseq reward model · SWE-bench Verifiedpass@1 (%): 39.2 -> 40.7Improved
BoostAPR vs GRPO under same backbone and dataQwen2.5-Coder-32B · SWE-bench Verifiedresolve rate (%): 36.1 -> 40.7Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy