CodeFort: Robust Training for Code Generation Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866929563697676288 |
|---|---|
| author | Zhang, Yuhao Wang, Shiqi Qian, Haifeng Wang, Zijian Shang, Mingyue Liu, Linbo Gouda, Sanjay Krishna Ray, Baishakhi Ramanathan, Murali Krishna Ma, Xiaofei Deoras, Anoop |
| author_facet | Zhang, Yuhao Wang, Shiqi Qian, Haifeng Wang, Zijian Shang, Mingyue Liu, Linbo Gouda, Sanjay Krishna Ray, Baishakhi Ramanathan, Murali Krishna Ma, Xiaofei Deoras, Anoop |
| contents | Code generation models are not robust to small perturbations, which often lead to incorrect generations and significantly degrade the performance of these models. Although improving the robustness of code generation models is crucial to enhancing user experience in real-world applications, existing research efforts do not address this issue. To fill this gap, we propose CodeFort, a framework to improve the robustness of code generation models, generalizing a large variety of code perturbations to enrich the training data and enabling various robust training strategies, mixing data augmentation, batch augmentation, adversarial logits pairing, and contrastive learning, all carefully designed to support high-throughput training. Extensive evaluations show that we increase the average robust pass rates of baseline CodeGen models from 14.79 to 21.74. We notably decrease the robustness drop rate from 95.02% to 54.95% against code-syntax perturbations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_01567 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | CodeFort: Robust Training for Code Generation Models Zhang, Yuhao Wang, Shiqi Qian, Haifeng Wang, Zijian Shang, Mingyue Liu, Linbo Gouda, Sanjay Krishna Ray, Baishakhi Ramanathan, Murali Krishna Ma, Xiaofei Deoras, Anoop Software Engineering Artificial Intelligence Code generation models are not robust to small perturbations, which often lead to incorrect generations and significantly degrade the performance of these models. Although improving the robustness of code generation models is crucial to enhancing user experience in real-world applications, existing research efforts do not address this issue. To fill this gap, we propose CodeFort, a framework to improve the robustness of code generation models, generalizing a large variety of code perturbations to enrich the training data and enabling various robust training strategies, mixing data augmentation, batch augmentation, adversarial logits pairing, and contrastive learning, all carefully designed to support high-throughput training. Extensive evaluations show that we increase the average robust pass rates of baseline CodeGen models from 14.79 to 21.74. We notably decrease the robustness drop rate from 95.02% to 54.95% against code-syntax perturbations. |
| title | CodeFort: Robust Training for Code Generation Models |
| topic | Software Engineering Artificial Intelligence |
| url | https://arxiv.org/abs/2405.01567 |