Research map / Large language models / Language, safety and interpretability
Do Not Have Enough Data? Deep Learning to the Rescue!
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 LAMBADA-generated, filtered examples | BERT classifier · ATIS, 5 samples per class | accuracy: 53.3 -> 75.7 | Helped |
| Add LAMBADA-generated, filtered examples | SVM classifier · ATIS, 5 samples per class | accuracy: 35.6 -> 56.5 | Helped |
| Add LAMBADA-generated, filtered examples | LSTM classifier · ATIS, 5 samples per class | accuracy: 29.0 -> 33.7 | Helped |
| Add LAMBADA-generated, filtered examples | BERT classifier · TREC, 5 samples per class | accuracy: 60.3 -> 64.3 | Helped |
| Add LAMBADA-generated, filtered examples | SVM classifier · TREC, 5 samples per class | accuracy: 42.7 -> 43.9 | Helped |
| Add LAMBADA-generated, filtered examples | LSTM classifier · TREC, 5 samples per class | accuracy: 17.7 -> 25.8 | Helped |
| Add LAMBADA-generated, filtered examples | BERT classifier · WVA, 5 samples per class | accuracy: 67.2 -> 68.6 | Helped |
| Add LAMBADA-generated, filtered examples | SVM classifier · WVA, 5 samples per class | accuracy: 60.2 -> 62.9 | Helped |
| Add LAMBADA-generated, filtered examples | LSTM classifier · WVA, 5 samples per class | accuracy: 26.0 -> 32.0 | Helped |
| Use LAMBADA rather than EDA for augmentation | SVM classifier · TREC, 5 samples per class | accuracy: 44.8 vs. 43.9 | No clear effect |
| Augment with LAMBADA rather than unlabeled-data weak labeling | BERT classifier · ATIS, 5 samples per class | accuracy: 54.5 -> 75.7 | Helped |
| Augment with LAMBADA rather than unlabeled-data weak labeling | SVM classifier · ATIS, 5 samples per class | accuracy: 23.5 -> 56.5 | Helped |
Closest papers by meaning
- ALP: Data Augmentation Using Lexicalized PCFGs for Few-Shot Text Classification (AAAI 2022 · similarity 0.87)
- Compositional Generalization for Multi-Label Text Classification: A Data-Augmentation Approach (AAAI 2024 · similarity 0.86)
- Not All LLM-Generated Data Are Equal: Rethinking Data Weighting in Text Classification (ICLR 2025 · similarity 0.85)
- Escaping Collapse: The Strength of Weak Data for Large Language Model Training (NeurIPS 2025 · similarity 0.84)
- RoPDA: Robust Prompt-Based Data Augmentation for Low-Resource Named Entity Recognition (AAAI 2024 · similarity 0.83)
- Curated LLM: Synergy of LLMs and Data Curation for tabular augmentation in low-data regimes (ICML 2024 · similarity 0.83)
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.