Research map / Label-efficient and robust learning / Recognition and long-tail
MetaTPTrans: A Meta Learning Approach for Multilingual Code Representation Learning
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 |
|---|---|---|---|
| Generate language-specific context-token projection weights (MetaTPTrans-α) rather than use a unified multilingual TPTrans | TPTrans · Code summarization: Python, Ruby, JavaScript, and Go | F1: 1.11, 1.47, 2.40 and 2.29 | Helped |
| Generate language-specific AST path-projection weights (MetaTPTrans-β) instead of context-token weights (ablation) | MetaTPTrans · Code summarization: Python, Ruby, JavaScript, and Go | F1 | Mixed |
| Generate language-specific weights for both context-token and AST path projections (MetaTPTrans-γ) (ablation) | MetaTPTrans · Code summarization: Python, Ruby, JavaScript, and Go | F1 | Mixed |
| Generate language-specific context-token projection weights (MetaTPTrans-α) rather than use unified multilingual TPTrans | TPTrans · Code completion: Python, Ruby, JavaScript, and Go | Top-1 accuracy: 7.32, 5.91, 6.07 and 7.02 | Helped |
| Generate language-specific context-token projection weights (MetaTPTrans-α) rather than use unified multilingual TPTrans | TPTrans · Code completion: Python, Ruby, JavaScript, and Go | Top-5 accuracy: 10.18, 11.95, 11.71 and 12.12 | Helped |
| Generate language-specific AST path-projection weights (MetaTPTrans-β) instead of context-token weights (ablation) | MetaTPTrans · Code completion: Python, Ruby, JavaScript, and Go | Top-1 accuracy | Mixed |
| Generate language-specific weights for both context-token and AST path projections (MetaTPTrans-γ) (ablation) | MetaTPTrans · Code completion: Python, Ruby, JavaScript, and Go | Top-1 accuracy | Mixed |
| Remove the pointer network (ablation) | MetaTPTrans code summarization models · Code summarization without pointer network | F1 | No clear effect |
Closest papers by meaning
- Language-Agnostic Representation Learning of Source Code from Structure and Context (ICLR 2021 · similarity 0.91)
- Multilingual Code Snippets Training for Program Translation (AAAI 2022 · similarity 0.87)
- CODE REPRESENTATION LEARNING AT SCALE (ICLR 2024 · similarity 0.87)
- CONSIDER: Commonalities and Specialties Driven Multilingual Code Retrieval Framework (AAAI 2024 · similarity 0.86)
- CoRNStack: High-Quality Contrastive Data for Better Code Retrieval and Reranking (ICLR 2025 · similarity 0.84)
- AdaCCD: Adaptive Semantic Contrasts Discovery Based Cross Lingual Adaptation for Code Clone Detection (AAAI 2024 · 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.