Learning Autocompletion from Real-World Datasets

2020/11/09 by Gareth Ari Aye, Aye, Gareth Ari, Seohyun Kim +3 · 1 voice
Computer Science · #Software Engineering Research #Software Reliability and Analysis Research #Software System Performance and Reliability #cs.SE

paper · pdf · doi:10.48550/arxiv.2011.04542

Abstract

Code completion is a popular software development tool integrated into all major IDEs. Many neural language models have achieved promising results in completion suggestion prediction on synthetic benchmarks. However, a recent study When Code Completion Fails: a Case Study on Real-World Completions demonstrates that these results may not translate to improvements in real-world performance. To combat this effect, we train models on real-world code completion examples and find that these models outperform models trained on committed source code and working version snapshots by 12.8% and 13.8% accuracy respectively. We observe this improvement across modeling technologies and show through A/B testing that it corresponds to a 6.2% increase in programmers' actual autocompletion usage. Furthermore, our study characterizes a large corpus of logged autocompletion usages to investigate why training on real-world examples leads to stronger models.

Discussions

• Learning Autocompletion from Real-World Datasets [hn, 3 points, 0 comments]

Related

• Improving Code Autocompletion with Transfer Learning
• Language Models for Code Completion: A Practical Evaluation
• Towards Full-line Code Completion with Neural Language Models
• Sequence Model Design for Code Completion in the Modern IDE
• When Neural Code Completion Models Size up the Situation: Attaining Cheaper and Faster Completion through Dynamic Model Inference