Research map / Architectures and efficiency / Attention and transformers
Guiding Attention in Sequence-to-Sequence Models for Dialogue Act Prediction
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 seq2seq instead of CRF with a GRU encoder | Dialogue act prediction · SwDA dev | accuracy: 77.7 -> 81.9 | Helped |
| Use seq2seq instead of CRF with a GRU encoder | Dialogue act prediction · MRDA dev | accuracy: 88.3 -> 88.5 | Helped |
| Use seq2seq instead of CRF with an HGRU encoder | Dialogue act prediction · SwDA dev | accuracy: 81.6 -> 82.4 | Helped |
| Use seq2seq instead of CRF with an HGRU encoder | Dialogue act prediction · MRDA dev | accuracy: 90.0 -> 90.0 | No clear effect |
| Add hierarchical word-and-utterance encoding instead of a vanilla encoder (ablation) | seq2seq · SwDA dev | accuracy: 81.6 -> 82.4 | Helped |
| Add hierarchical word-and-utterance encoding instead of a vanilla encoder (ablation) | seq2seq · MRDA dev | accuracy: 88.5 -> 90.0 | Helped |
| Add hard guided attention rather than no attention (ablation) | seq2seq + HGRU · SwDA dev | accuracy: 82.4 -> 84.0 | Helped |
| Add hard guided attention rather than no attention (ablation) | seq2seq + HGRU · MRDA dev | accuracy: 90.0 -> 90.2 | Helped |
| Replace hard guided attention with vanilla attention (ablation) | seq2seq + HGRU · SwDA dev | accuracy: 84.0 -> 82.3 | Hurt |
| Replace hard guided attention with vanilla attention (ablation) | seq2seq + HGRU · MRDA dev | accuracy: 90.2 -> 89.9 | Hurt |
| Add a persona layer to the hierarchical encoder (ablation) | seq2seq + hard guided attention · SwDA dev | accuracy: 84.0 -> 83.5 | Hurt |
| Add a persona layer to the hierarchical encoder (ablation) | seq2seq + hard guided attention · MRDA dev | accuracy: 90.2 -> 86.9 | Hurt |
Closest papers by meaning
- Dialogue Act Sequence Labeling Using Hierarchical Encoder With CRF (AAAI 2018 · similarity 0.89)
- A Knowledge-Grounded Neural Conversation Model (AAAI 2018 · similarity 0.89)
- Mask & Focus: Conversation Modelling by Learning Concepts (AAAI 2020 · similarity 0.87)
- TA&AT: Enhancing Task-Oriented Dialog with Turn-Level Auxiliary Tasks and Action-Tree Based Scheduled Sampling (AAAI 2024 · similarity 0.86)
- Structure-Aware Multimodal Sequential Learning for Visual Dialog (AAAI 2024 · similarity 0.85)
- Learning to Clarify: Multi-turn Conversations with Action-Based Contrastive Self-Training (ICLR 2025 · similarity 0.85)
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.