mlsanityData to 7 Oct 2026

Self-training for few-shot NLP

28 papers0 this weekHigh confidence in this grouping: the papers form one clear topicHigh confidence in this grouping: the papers form one clear topic · About

Keywords: self-training, few-shot, unlabeled, classification, language

Key facts. Self-training for few-shot NLP: 28 papers on arXiv since 2021 (data through 7 Oct 2026). 4 of its 28 mapped papers (14%) cite “STraTA: Self-Training with Task Augmentation for Better Few-shot Learning” (2021, arXiv 2109.06270). 2 of its 28 mapped papers (7%) cite “Self-training with Few-shot Rationalization: Teacher Explanations Aid Student in Few-shot NLU” (2021, arXiv 2109.08259).

What’s new

Nothing new since 17 Sep. Latest: 20 Jul 2026.

Foundations

Most cited by papers here

  1. STraTA: Self-Training with Task Augmentation for Better Few-shot Learning2021cited by 4 here
  2. Self-training with Few-shot Rationalization: Teacher Explanations Aid Student in Few-shot NLU2021cited by 2 here
  3. Revisiting Self-Training for Few-Shot Learning of Language Model2021cited by 2 here
  4. LiST: Lite Prompted Self-training Makes Parameter-Efficient Few-shot Learners2021cited by 2 here
  5. Zero-Shot Text Classification with Self-Training2022cited by 2 here

Lineage

The papers this direction builds on, oldest first.

  1. 2021STraTA: Self-Training with Task Augmentation for Better Few-shot Learningcited by 4 here
  2. 2021Self-training with Few-shot Rationalization: Teacher Explanations Aid Student in Few-shot NLUcited by 2 here
  3. 2021Revisiting Self-Training for Few-Shot Learning of Language Modelcited by 2 here
  4. 2021LiST: Lite Prompted Self-training Makes Parameter-Efficient Few-shot Learnerscited by 2 here
  5. 2022STaR: Bootstrapping Reasoning With Reasoningcited by 2 here
  6. 2022Self-Generated In-Context Learning: Leveraging Auto-regressive Language Models as a Demonstration Generatorcited by 1 here
  7. 2022Zero-Shot Text Classification with Self-Trainingcited by 2 here
  8. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2 here
  9. 2023Self-ICL: Zero-Shot In-Context Learning with Self-Generated Demonstrationscited by 1 here
  10. 2023Taming Self-Training for Open-Vocabulary Object Detectioncited by 1 here
  11. 2023MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Modelscited by 1 here
  12. 2023DST-Det: Simple Dynamic Self-Training for Open-Vocabulary Object Detectioncited by 1 here
6 more
  1. 2023DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelinescited by 1 here
  2. 2023Synthetic Data Generation with Large Language Models for Text Classification: Potential and Limitationscited by 1 here
  3. 2023Self-Improving for Zero-Shot Named Entity Recognition with Large Language Modelscited by 1 here
  4. 2024Key-Point-Driven Data Synthesis with its Enhancement on Mathematical Reasoningcited by 1 here
  5. 2024On LLMs-Driven Synthetic Data Generation, Curation, and Evaluation: A Surveycited by 1 here
  6. 2024Scaling Synthetic Data Creation with 1,000,000,000 Personascited by 1 here

Browse

23 of 28, ranked 6–28

  1. 6Self-ICL: Zero-Shot In-Context Learning with Self-Generated Demonstrations20231
  2. 7Self-Improving for Zero-Shot Named Entity Recognition with Large Language Models20231
  3. 8Self-Generated In-Context Learning: Leveraging Auto-regressive Language Models as a Demonstration Generator20221
  4. 9A Classifier That Teaches Itself: Self-Improving, Frozen-gate Training (SIFT) for Dynamic Document Classification20260
  5. 10From Few to Many: Self-Improving Many-Shot Reasoners Through Iterative Optimization and Generation20250
  6. 11Keyword-Centric Prompting for One-Shot Event Detection with Self-Generated Rationale Enhancements20250
  7. 12Multi-Agent VLMs Guided Self-Training with PNU Loss for Low-Resource Offensive Content Detection20250
  8. 13Enhancing Low-Resource LLMs Classification with PEFT and Synthetic Data20240
  9. 14Self-Training for Sample-Efficient Active Learning for Text Classification with Pre-Trained Language Models20240
  10. 15Self-Training with Pseudo-Label Scorer for Aspect Sentiment Quad Prediction20240
All 23
  1. 16ReverseNER: A Self-Generated Example-Driven Framework for Zero-Shot Named Entity Recognition with Large Language Models20240
  2. 17DS2-ABSA: Dual-Stream Data Synthesis with Label Refinement for Few-Shot Aspect-Based Sentiment Analysis20240
  3. 18Generation-driven Contrastive Self-training for Zero-shot Text Classification with Instruction-following LLM20230
  4. 19Accelerating Neural Self-Improvement via Bootstrapping20230
  5. 20Few Shot Rationale Generation using Self-Training with Dual Teachers20230
  6. 21KEST: Kernel Distance Based Efficient Self-Training for Improving Controllable Text Generation20230
  7. 22Self-training Strategies for Sentiment Analysis: An Empirical Study20230
  8. 23Uncertainty-aware Parameter-Efficient Self-training for Semi-supervised Language Understanding20230
  9. 24Boosting Prompt-Based Self-Training With Mapping-Free Automatic Verbalizer for Multi-Class Classification20230
  10. 25Neural Networks Against (and For) Self-Training: Classification with Small Labeled and Large Unlabeled Sets20230
  11. 26Self-Training Pre-Trained Language Models for Zero- and Few-Shot Multi-Dialectal Arabic Sequence Labeling20210
  12. 27Self-training Improves Pre-training for Few-shot Learning in Task-oriented Dialog Systems20210
  13. 28AcTune: Uncertainty-aware Active Self-Training for Semi-Supervised Active Learning with Pretrained Language Models20210