Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Meta-Transfer Learning for Low-Resource 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 |
|---|---|---|---|
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · AESLC | ROUGE-2: 4.91 -> 10.79 | Helped |
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · BillSum | ROUGE-2: 18.49 -> 18.61 | Helped |
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · Gigaword | ROUGE-2: 10.04 -> 11.86 | Helped |
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · Multi-News | ROUGE-2: 12.56 -> 12.78 | Helped |
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · NEWSROOM | ROUGE-2: 17.78 -> 25.40 | Helped |
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · Reddit-TIFU | ROUGE-2: 5.16 -> 6.41 | Helped |
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · arXiv | ROUGE-2: 8.16 -> 10.26 | Helped |
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · PubMed | ROUGE-2: 10.58 -> 10.05 | Hurt |
| Meta-learn adapter initialization across source corpora, then adapt with 10 target examples | MTL-ABS · WikiHow | ROUGE-2: 6.54 -> 8.16 | Helped |
| Restrict learning to adapter modules and layer normalization rather than fine-tuning the full model (ablation) | MTL-ABS · AESLC | ROUGE-2: 7.20 -> 10.79 | Helped |
| Restrict learning to adapter modules and layer normalization rather than fine-tuning the full model (ablation) | MTL-ABS · Gigaword | ROUGE-2: 8.89 -> 11.86 | Helped |
| Restrict learning to adapter modules and layer normalization rather than fine-tuning the full model (ablation) | MTL-ABS · Reddit-TIFU | ROUGE-2: 4.00 -> 6.41 | Helped |
Closest papers by meaning
- Controlling the Amount of Verbatim Copying in Abstractive Summarization (AAAI 2020 · similarity 0.89)
- MeanSum: A Neural Model for Unsupervised Multi-Document Abstractive Summarization (ICML 2019 · similarity 0.88)
- MultiSumm: Towards a Unified Model for Multi-Lingual Abstractive Summarization (AAAI 2020 · similarity 0.88)
- Joint Parsing and Generation for Abstractive Summarization (AAAI 2020 · similarity 0.87)
- Distilling Structured Rationale from Large Language Models to Small Language Models for Abstractive Summarization (AAAI 2025 · similarity 0.87)
- Learning from Failure: Improving Meeting Summarization without Good Samples (AAAI 2024 · similarity 0.86)
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.