Research map / Label-efficient and robust learning / Recognition and long-tail
Discovering New Intents with Deep Aligned Clustering
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 |
|---|---|---|---|
| Add aligned pseudo-labels instead of reinitializing the classifier, with pre-training (ablation) | DeepAligned · CLINC | ACC: 55.80 -> 86.49 | Helped |
| Add aligned pseudo-labels instead of reinitializing the classifier, with pre-training (ablation) | DeepAligned · BANKING | ACC: 41.32 -> 64.90 | Helped |
| Add aligned pseudo-labels instead of reinitializing the classifier, without pre-training (ablation) | DeepAligned · CLINC | ACC: 23.02 -> 28.63 | Helped |
| Add aligned pseudo-labels instead of reinitializing the classifier, without pre-training (ablation) | DeepAligned · BANKING | ACC: 13.67 -> 15.42 | Helped |
| Add labeled-data pre-training while using aligned pseudo-labels (ablation) | DeepAligned · CLINC | ACC: 28.63 -> 86.49 | Helped |
| Add labeled-data pre-training while using aligned pseudo-labels (ablation) | DeepAligned · BANKING | ACC: 15.42 -> 64.90 | Helped |
| Use DeepAligned rather than BERT-DTC | intent clustering · CLINC | ACC: 74.15 -> 86.49 | Helped |
| Use DeepAligned rather than BERT-MCL | intent clustering · BANKING | ARI: 47.43 -> 53.64 | Helped |
| Estimate cluster count by eliminating low-confidence clusters using pre-trained features | DeepAligned · CLINC, 25% known classes | K error: 37.33 -> 18.67 | Helped |
| Estimate cluster count by eliminating low-confidence clusters using pre-trained features | DeepAligned · BANKING, 25% known classes | K error: 51.95 -> 14.29 | Helped |
| Estimate cluster count by eliminating low-confidence clusters using pre-trained features | DeepAligned · CLINC, 50% known classes | K error: 12.67 -> 13.33 | Hurt |
| Estimate cluster count by eliminating low-confidence clusters using pre-trained features | DeepAligned · BANKING, 50% known classes | K error: 7.79 -> 16.88 | Hurt |
Closest papers by meaning
- MIntRec2.0: A Large-scale Benchmark Dataset for Multimodal Intent Recognition and Out-of-scope Detection in Conversations (ICLR 2024 · similarity 0.86)
- End-to-end Learnable Clustering for Intent Learning in Recommendation (NeurIPS 2024 · similarity 0.85)
- Aligner²: Enhancing Joint Multiple Intent Detection and Slot Filling via Adjustive and Forced Cross-Task Alignment (AAAI 2024 · similarity 0.85)
- Multi-Granularity Open Intent Classification via Adaptive Granular-Ball Decision Boundary (AAAI 2025 · similarity 0.84)
- Uni-MIS: United Multiple Intent Spoken Language Understanding via Multi-View Intent-Slot Interaction (AAAI 2024 · similarity 0.83)
- Divide-Solve-Combine: An Interpretable and Accurate Prompting Framework for Zero-shot Multi-Intent Detection (AAAI 2025 · similarity 0.81)
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.