Research map / Large language models / Language, safety and interpretability
MultiSumm: Towards a Unified Model for Multi-Lingual Abstractive Summarization
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 auto-encoder training (ablation) | MultiSumm · Bosnian summarization | ROUGE-1: 21.01 -> 22.47 | Helped |
| add auto-encoder training (ablation) | MultiSumm · German summarization (SWISS) | ROUGE-1: 40.73 -> 43.41 | Helped |
| add translation and back-translation training (ablation) | MultiSumm · Croatian summarization | ROUGE-1: 21.62 -> 23.04 | Helped |
| add language model pretraining (ablation) | MultiSumm · Croatian summarization | ROUGE-1: 22.47 -> 23.04 | No clear effect |
| use fully shared encoder/decoder instead of partially shared (ablation) | MultiSumm · Croatian summarization | ROUGE-1: 22.78 -> 23.04 | No clear effect |
| add pretraining (LM + auto-encoder) to individual monolingual model | Individual transformer · Bosnian summarization | ROUGE-1: 13.76 -> 18.22 | Helped |
| add pretraining (LM + auto-encoder) to individual monolingual model | Individual transformer · Croatian summarization | ROUGE-1: 12.61 -> 18.56 | Helped |
| use multilingual training without pretraining instead of individual models | Multi-baseline · German summarization (SWISS) | ROUGE-1: 34.30 -> 39.33 | Mixed |
| use multilingual baseline instead of individual on rich-resource English | Multi-baseline · English summarization (Gigaword) | ROUGE-1: 36.54 -> 35.06 | Hurt |
| add multilingual learning stage to multilingual baseline | MultiSumm · Bosnian summarization | ROUGE-1: 20.27 -> 22.47 | Helped |
| MultiSumm vs individual + pretraining | MultiSumm · German summarization (SWISS) | ROUGE-1: 45.17 -> 43.41 | Hurt |
| MultiSumm vs individual + pretraining | MultiSumm · Croatian summarization | ROUGE-1: 18.56 -> 23.04 | Helped |
Closest papers by meaning
- Distilling Structured Rationale from Large Language Models to Small Language Models for Abstractive Summarization (AAAI 2025 · similarity 0.89)
- PersonalSum: A User-Subjective Guided Personalized Summarization Dataset for Large Language Models (NeurIPS 2024 · similarity 0.88)
- Language Models of Code Are Few-Shot Planners and Reasoners for Multi-Document Summarization with Attribution (AAAI 2025 · similarity 0.87)
- CAR-Transformer: Cross-Attention Reinforcement Transformer for Cross-Lingual Summarization (AAAI 2024 · similarity 0.87)
- Learning from Failure: Improving Meeting Summarization without Good Samples (AAAI 2024 · similarity 0.86)
- Scaling Up Video Summarization Pretraining with Large Language Models (CVPR 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.