Few-Shot RL & Meta-Learning Reading List

Curated by Mouhssine Rifaki | Stanford Electrical Engineering | Last updated August 2026

Few-shot RL aims to adapt to new tasks from a handful of episodes. This list traces gradient-based meta-learning, context-based adaptation, and modern few-shot policy learners.

Few-Shot RL & Meta-Learning: 10 key papers

1. On First-Order Meta-Learning Algorithms
   Nichol et al. arXiv 2018.
2. Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML
   Raghu et al. arXiv 2019.
3. Alpha MAML: Adaptive Model-Agnostic Meta-Learning
   Behl et al. arXiv 2019.
4. Recurrent Hypernetworks are Surprisingly Strong in Meta-RL
   Beck et al. arXiv 2023.
5. Meta-Learning with Latent Embedding Optimization
   Rusu et al. arXiv 2018.
6. Some Considerations on Learning to Explore via Meta-Reinforcement Learning
   Stadie et al. arXiv 2018.
7. Meta-Reinforcement Learning of Structured Exploration Strategies
   Gupta et al. arXiv 2018.
8. Learning to reinforcement learn
   Wang et al. arXiv 2016.
9. Meta-Learning without Memorization
   Yin et al. arXiv 2019.
10. Prototypical Networks for Few-shot Learning
    Snell et al. arXiv 2017.

Related reading lists

• Meta-Reinforcement Learning
• Memory-Augmented RL & Recurrent Policies
• Exploration in Reinforcement Learning
• Prompt Optimization & In-Context Learning for LLMs
• Neuroscience-Inspired Reinforcement Learning
• Retrieval-Augmented Generation
• Zero-Shot RL & Generalization
• Reward Learning & Reward Modeling