Research map / Label-efficient and robust learning / Knowledge distillation
Improving Neural Cross-Lingual Abstractive Summarization via Employing Optimal Transport Distance for Knowledge Distillation
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 |
|---|---|---|---|
| Distill monolingual teacher representations into the cross-lingual student using Sinkhorn divergence | Cross-lingual summarization · En2Zh full-dataset | ROUGE-1: 44.16 -> 44.75 | Helped |
| Distill monolingual teacher representations into the cross-lingual student using Sinkhorn divergence | Cross-lingual summarization · Zh2En full-dataset | ROUGE-1: 35.65 -> 36.93 | Helped |
| Distill monolingual teacher representations into the cross-lingual student using Sinkhorn divergence | Cross-lingual summarization · En2Ar full-dataset | ROUGE-L: 27.25 -> 32.38 | Helped |
| Distill monolingual teacher representations into the cross-lingual student using Sinkhorn divergence | Cross-lingual summarization · Zh2En minimum | ROUGE-1: 21.03 -> 22.37 | Helped |
| Use mean-pooling with cosine similarity instead of Sinkhorn divergence (ablation) | Knowledge-distilled cross-lingual summarization · En2Zh full-dataset | ROUGE-1: 44.75 -> 44.20 | Hurt |
| Use mean-pooling with mean squared error instead of Sinkhorn divergence (ablation) | Knowledge-distilled cross-lingual summarization · En2Zh full-dataset | ROUGE-1: 44.75 -> 44.14 | Hurt |
| Use max-pooling with cosine similarity instead of Sinkhorn divergence (ablation) | Knowledge-distilled cross-lingual summarization · En2Zh full-dataset | ROUGE-1: 44.75 -> 44.29 | Hurt |
| Use max-pooling with mean squared error instead of Sinkhorn divergence (ablation) | Knowledge-distilled cross-lingual summarization · En2Zh full-dataset | ROUGE-1: 44.75 -> 44.23 | Hurt |
Closest papers by meaning
- CAR-Transformer: Cross-Attention Reinforcement Transformer for Cross-Lingual Summarization (AAAI 2024 · similarity 0.89)
- MultiSumm: Towards a Unified Model for Multi-Lingual Abstractive Summarization (AAAI 2020 · similarity 0.88)
- Multi-Level Optimal Transport for Universal Cross-Tokenizer Knowledge Distillation on Language Models (AAAI 2025 · similarity 0.86)
- Soft Token Alignment for Cross-Lingual Reasoning (NeurIPS 2026 · similarity 0.85)
- Distilling Structured Rationale from Large Language Models to Small Language Models for Abstractive Summarization (AAAI 2025 · similarity 0.85)
- Order-Level Attention Similarity Across Language Models: A Latent Commonality (NeurIPS 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.