P-NPR: Practical Neural Program Repair via Learning to Ensemble
- 28 April 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Software Engineering
- Vol. PP (00985589) , 1-16
Abstract
The rapid emergence of Large Language Model (LLM)-based program repair tools has significantly advanced the potential of automated program repair. Nevertheless, traditional Neural Program Repair (NPR) tools continue to offer unique value in resource-constrained or latency-sensitive scenarios due to their lightweight nature. Over 50 such NPR tools have been designed and proven effective at fixing various bugs using only the buggy code as input. Recent studies further demonstrate that combining these tools through ensemble strategies can enhance overall repair performance by leveraging their complementary strengths. However, existing ensemble approaches still suffer from two critical limitations: (1) they rely on manual feature summarization, requiring labor-intensive labeling of bug characteristics, and (2) they ignore the common “all-tools-fail” scenario (e.g., even a set of 15 state-of-the-art tools leaves over 66% of Defects4J bugs unfixed), which incurs unnecessary computational overhead. To address these issues, we propose Practical Neural Program Repair (P-NPR), a novel ensemble strategy that employs multilabel classification to automatically learn to characterize bugs and introduces a dedicated “AllFailure” label to explicitly sidestep futile repair attempts. Experimental results show that P-NPR outperforms the state-of-the-art ensemble strategy (i.e., P-EPR) by over 65% in Top1-accuracy, while reducing runtime by up to 45% and GPU memory usage by up to 23% across evaluated datasets.Keywords
Funding Information
- National Natural Science Foundation of China (62572235)
- Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China (JYB2025XDXM118)
- CCF-Huawei Populus Grove Fund
- CCF-Huawei Populus Grove Fund 111 Center (B26023)
This publication has 39 references indexed in Scilit:
- DALO-APR: LLM-based automatic program repair with data augmentation and loss function optimizationThe Journal of Supercomputing, 2025
- Gamma: Revisiting Template-Based Automated Program Repair Via Mask Prediction2023
- Tare: Type-Aware Neural Program Repair2023
- Automated Program Repair in the Era of Large Pre-trained Language Models2023
- Template-based Neural Program Repair2023
- KNOD: Domain Knowledge Distilled Tree Decoder for Automated Program Repair2023
- E-APR: Mapping the effectiveness of automated program repair techniquesEmpirical Software Engineering, 2021
- CURE: Code-Aware Neural Machine Translation for Automatic Program Repair2021
- Pre-trained models for natural language processing: A surveyScience China Technological Sciences, 2020
- Identification of common molecular subsequencesJournal of Molecular Biology, 1981