Research map / Architectures and efficiency / Attention and transformers
A Grammar-Based Structural CNN Decoder for Code Generation
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 |
|---|---|---|---|
| replace the pre-order CNN with an LSTM (ablation) | Grammar-based structural CNN decoder · HearthStone | Acc+: 30.3 -> 21.2 | Hurt |
| replace the pre-order CNN with an LSTM (ablation) | Grammar-based structural CNN decoder · HearthStone | BLEU: 79.6 -> 78.8 | Hurt |
| add the predicted rule CNN (ablation) | Grammar-based structural CNN decoder · HearthStone | Acc+: 24.2 -> 30.3 | Helped |
| add the pre-order CNN (ablation) | Grammar-based structural CNN decoder · HearthStone | Acc+: 25.8 -> 30.3 | Helped |
| add the pre-order CNN (ablation) | Grammar-based structural CNN decoder · HearthStone | BLEU: 80.4 -> 79.6 | Hurt |
| add the tree-based CNN (ablation) | Grammar-based structural CNN decoder · HearthStone | Acc+: 25.8 -> 30.3 | Helped |
| add the tree-path CNN (ablation) | Grammar-based structural CNN decoder · HearthStone | Acc+: 28.8 -> 30.3 | Helped |
| add attentive pooling (ablation) | Grammar-based structural CNN decoder · HearthStone | Acc+: 24.2 -> 30.3 | Helped |
| add the scope name controller (ablation) | Grammar-based structural CNN decoder · HearthStone | Acc+: 25.8 -> 30.3 | Helped |
Closest papers by meaning
- Convolutional Neural Networks over Tree Structures for Programming Language Processing (AAAI 2016 · similarity 0.89)
- TreeGen: A Tree-Based Transformer Architecture for Code Generation (AAAI 2020 · similarity 0.87)
- Towards a theory of how the structure of language is acquired by deep neural networks (NeurIPS 2024 · similarity 0.85)
- Grammar-Aligned Decoding (NeurIPS 2024 · similarity 0.85)
- Diffusion On Syntax Trees For Program Synthesis (ICLR 2025 · similarity 0.84)
- CODE REPRESENTATION LEARNING AT SCALE (ICLR 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.