Research map / Architectures and efficiency / Attention and transformers
Tree-to-tree Neural Networks for Program Translation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use tree-to-tree translation instead of the strongest listed baseline | Java-to-C# translation · Lucene, POI, Itext, JGit, JTS | program accuracy: 72.8% vs 40.0%; 72.2% vs 48.2%; 67.5% vs 40.6%; 68.7% vs 48.5%; 68.2% vs 26.3% | Helped |
| Use tree-to-tree translation instead of mppSMT | Java-to-C# translation · Antlr | program accuracy: 31.9% vs 49.1% | Hurt |
| Train using data from the other 5 projects | Tree2tree · Antlr | program accuracy: 31.9% -> 58.3% | Helped |
| Use Tree2tree instead of the strongest listed neural baseline | CoffeeScript-JavaScript translation · CJ-AL, CJ-BL, JC-BL | program accuracy: 97.15% vs 80.82%; 95.60% vs 83.61%; 75.62% vs 74.51% | Helped |
| Use Tree2tree instead of the strongest listed neural baseline | Synthetic translation · SYN-L | program accuracy: 97.50% vs 68.11% | Helped |
| Add parent attention feeding (ablation) | Tree2tree · JC-BL | program accuracy: 44.40% -> 75.62% | Helped |
| Add attention (ablation) | Tree2tree · JC-BL | program accuracy: 0% -> 75.62% | Helped |
| Add parent attention feeding (ablation) | Tree2tree · SYN-L | program accuracy: 57.42% -> 97.50% | Helped |
| Add attention (ablation) | Tree2tree · SYN-L | program accuracy: 0% -> 97.50% | Helped |
Closest papers by meaning
- Convolutional Neural Networks over Tree Structures for Programming Language Processing (AAAI 2016 · similarity 0.88)
- TreeGen: A Tree-Based Transformer Architecture for Code Generation (AAAI 2020 · similarity 0.87)
- Multilingual Code Snippets Training for Program Translation (AAAI 2022 · similarity 0.87)
- Diffusion On Syntax Trees For Program Synthesis (ICLR 2025 · similarity 0.84)
- Guess & Sketch: Language Model Guided Transpilation (ICLR 2024 · similarity 0.84)
- QiMeng-MuPa: Mutual-Supervised Learning for Sequential-to-Parallel Code Translation (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.