Metalearning Lab: 50+ Algorithms and Benchmarks for Few-Shot Learning

Metalearning Lab is a repository of code, methods, and dataset references for meta-learning research curated by WangJingyao07 and updated in March 2024. It serves as the technical foundation for the Awesome-META+ platform, aggregating resources for few-shot learning, meta-reinforcement learning, and meta-optimization.

Use Cases

• Benchmarking few-shot-learning models using the provided algorithm implementations.
• Implementing meta-rl (meta-reinforcement learning) environments based on the repository's methods.
• Researching metaoptimization techniques through the curated collection of academic papers.

Strengths

• Includes meta-RL and meta-optimization tags for specialized research
• Updated in March 2024
• Directly supports the Awesome-META+ platform

Limitations

• Undefined file formats and sizes
• Lack of a unified data schema
• Requires external navigation to access some linked datasets

Provenance

Source

WangJingyao07 (GitHub)

Collection Method

Curated from academic literature and existing meta-learning benchmarks.

Freshness

Last updated March 2024.