Research map / Label-efficient and robust learning / Recognition and long-tail
DARE: Disentanglement-Augmented Rationale Extraction
Linan Yue, Qi Liu, Yichao Du, Yanqing An, Li Wang, Enhong Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 disentanglement of non-rationale tokens with MI minimization (HardKuma -> DARE) (ablation) | HardKuma · BeerAdvocate Appearance | token F1: 78.3 -> 82.9 | Improved |
| Add disentanglement of non-rationale tokens with MI minimization (HardKuma -> DARE) (ablation) | HardKuma · BeerAdvocate Smell | token F1: 47.5 -> 61.2 | Improved |
| Add disentanglement of non-rationale tokens with MI minimization (HardKuma -> DARE) (ablation) | HardKuma · BeerAdvocate Palate | token F1: 63.1 -> 69.9 | Improved |
| Replace CLUB_NCE with L1Out as MI minimizer (ablation) | DARE · BeerAdvocate Appearance | token F1: 82.9 -> 41.3 | Hurt |
| Replace CLUB_NCE with CLUB as MI minimizer (ablation) | DARE · BeerAdvocate Smell | token F1: 61.2 -> 58.3 | Hurt |
| Replace CLUB_NCE with CLUB as MI minimizer (ablation) | DARE · BeerAdvocate Palate | token F1: 69.9 -> 67.4 | Hurt |
| Replace CLUB_NCE with L1Out as MI minimizer (ablation) | DARE · CAIL2018 Terms of Penalty | macro-F1: 30.65 -> 3.72 | Hurt |
| Replace CLUB_NCE with CLUB as MI minimizer (ablation) | DARE · CAIL2018 Charges | macro-F1: 84.36 -> 74.84 | Hurt |
| Add reconstruction loss to DARE (ablation) | DARE · BeerAdvocate Appearance | token F1: 82.9 -> 82.6 | No clear effect |
| Replace RCNN with BigBird in DARE versus HardKuma | DARE · BeerAdvocate Appearance | token F1: 84.4 -> 87.5 | Improved |
| Use DARE instead of InfoCal(HK) with external guider | InfoCal(HK) · ERASER Movie Reviews | token F1: 33.2 -> 33.4 | Improved |
| Use DARE instead of HardKuma | HardKuma · Human evaluation, charge prediction | usefulness: 3.95 -> 4.58 | Improved |
Closest papers by meaning
- Breaking Free from MMI: A New Frontier in Rationalization by Probing Input Utilization (ICLR 2025 · similarity 0.87)
- Is the MMI Criterion Necessary for Interpretability? Degenerating Non-causal Features to Plain Noise for Self-Rationalization (NeurIPS 2024 · similarity 0.83)
- Towards Faithful Explanations: Boosting Rationalization with Shortcuts Discovery (ICLR 2024 · similarity 0.83)
- Learning Robust Rationales for Model Explainability: A Guidance-Based Approach (AAAI 2024 · similarity 0.83)
- $\alpha$TC-VAE: On the relationship between Disentanglement and Diversity (ICLR 2024 · similarity 0.81)
- Rethinking Dimensional Rationale in Graph Contrastive Learning from Causal Perspective (AAAI 2024 · 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.
