Research map / Architectures and efficiency / Attention and transformers
Mask & Focus: Conversation Modelling by Learning Concepts
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 Mask & Focus instead of MrRNN-Noun | Ubuntu Dialogue Corpus · Ubuntu Dialogue Corpus | Entity F1: 6.31 vs 7.82 | Helped |
| Use Mask & Focus instead of MrRNN-ActEnt | Ubuntu Dialogue Corpus · Ubuntu Dialogue Corpus | Activity F1: 11.43 vs 5.43 | Hurt |
| Use Mask & Focus instead of HRED-attend | Ubuntu Dialogue Corpus · Ubuntu Dialogue Corpus | Tense Acc.: 29.75 vs 24.47 | Hurt |
| Use Mask & Focus instead of HRED-attend | Ubuntu Dialogue Corpus · Ubuntu Dialogue Corpus | Cmd Acc.: 99.89 vs 93.19 | Hurt |
| Use Mask & Focus instead of MrRNN-Noun | Ubuntu Dialogue Corpus human evaluation · Ubuntu Dialogue Corpus | Relevance: 1.56 vs 1.79 | Helped |
| Use Mask & Focus instead of MrRNN-Noun | Ubuntu Dialogue Corpus human evaluation · Ubuntu Dialogue Corpus | GR Ratio: 0.32 vs 0.18 | Helped |
| Train on fewer examples using Mask & Focus rather than HRED | Tech Support · Tech Support | BLEU | Helped |
| Increase training data size | Mask & Focus versus HRED on Tech Support · Tech Support | BLEU | Mixed |
Closest papers by meaning
- DialogBERT: Discourse-Aware Response Generation via Learning to Recover and Rank Utterances (AAAI 2021 · similarity 0.87)
- Guiding Attention in Sequence-to-Sequence Models for Dialogue Act Prediction (AAAI 2020 · similarity 0.87)
- AvgOut: A Simple Output-Probability Measure to Eliminate Dull Responses (AAAI 2020 · similarity 0.87)
- Compressed Context Memory for Online Language Model Interaction (ICLR 2024 · similarity 0.85)
- Generative Adapter: Contextualizing Language Models in Parameters with A Single Forward Pass (ICLR 2025 · similarity 0.85)
- Talking Heads: Understanding Inter-Layer Communication in Transformer Language Models (NeurIPS 2024 · 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.