2018/12/04 by Tatsunori Hashimoto, Tatsunori B. Hashimoto, Hashimoto, Tatsunori B. +6 · 102 citations
Computer Science · Mathematics · #Advanced Malware Detection Techniques #Artificial intelligence #Benchmark (surveying) #Code (set theory) #Computer science #Data mining #Edit distance #FOS: Computer and information sciences #Information retrieval #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine learning #Metric (unit) #Natural language processing #Programming language #Python (programming language) #Scratch #Software Engineering Research #Source code #Task (project management) #Web Data Mining and Analysis #cs.LG #stat.ML
paper · pdf · doi:10.48550/arxiv.1812.01194
published in arXiv (Cornell University) 31, 10052-10062 (Cornell University) · To appear, NeurIPS 2018
arxiv created 2018/12/04 · openalex publication_date 2018/12/04 · arxiv updated 2018/12/05 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
For the task of generating complex outputs such as source code, editing existing outputs can be easier than generating complex outputs from scratch. With this motivation, we propose an approach that first retrieves a training example based on the input (e.g., natural language description) and then edits it to the desired output (e.g., code). Our contribution is a computationally efficient method for learning a retrieval model that embeds the input in a task-dependent way without relying on a hand-crafted metric or incurring the expense of jointly training the retriever with the editor. Our retrieve-and-edit framework can be applied on top of any base model. We show that on a new autocomplete task for GitHub Python code and the Hearthstone cards benchmark, retrieve-and-edit significantly boosts the performance of a vanilla sequence-to-sequence model on both tasks.