Code Completion
Code completion automates the suggestion of code snippets during software development, aiming to boost developer productivity and code quality. Current research emphasizes improving the accuracy and efficiency of these tools, particularly focusing on leveraging large language models (LLMs) and incorporating repository-level context through techniques like retrieval augmentation and graph-based approaches. This field is crucial for enhancing software development workflows, with ongoing efforts addressing challenges such as security vulnerabilities in LLM-based tools and the need for more realistic and comprehensive evaluation benchmarks.
Papers
Measuring memorization in RLHF for code completion
Aneesh Pappu, Billy Porter, Ilia Shumailov, Jamie Hayes
Long Code Arena: a Set of Benchmarks for Long-Context Code Models
Egor Bogomolov, Aleksandra Eliseeva, Timur Galimzyanov, Evgeniy Glukhov, Anton Shapkin, Maria Tigina, Yaroslav Golubev, Alexander Kovrigin, Arie van Deursen, Maliheh Izadi, Timofey Bryksin
GitHub Copilot: the perfect Code compLeeter?
Ilja Siroš, Dave Singelée, Bart Preneel