This repository is a curated collection of resources on predictive uncertainty estimation in deep learning, targeting researchers and practitioners in AI. It aims to provide a comprehensive overview of the field by organizing papers, datasets, libraries, and tutorials, facilitating the understanding and application of uncertainty quantification techniques.

How It Works

The repository categorizes resources by methodology (e.g., Bayesian methods, ensembles, post-hoc methods, output-space modeling), application domain (e.g., classification, regression, anomaly detection), and resource type (papers, datasets, libraries). This structured approach allows users to navigate the vast landscape of uncertainty estimation, discover relevant research, and find practical tools.

Quick Start & Requirements

This is a curated list of resources, not a runnable software package. No installation or specific requirements are needed to browse the content.

Highlighted Details

• Extensive categorization of papers by specific uncertainty estimation techniques and application areas.
• Links to numerous datasets and benchmarks for evaluating uncertainty methods.
• Curated list of popular libraries and frameworks (PyTorch, TensorFlow, JAX) for implementing uncertainty estimation.
• Includes links to lectures, tutorials, and books for learning and deeper understanding.

Maintenance & Community

The repository is marked with the "Awesome" badge, indicating community curation. Contributions are welcomed via pull requests or GitHub discussions for adding missing papers or suggesting improvements.

Licensing & Compatibility

The repository itself is licensed under the MIT License, allowing for broad use and modification. Individual resources (papers, code) retain their original licenses.

Limitations & Caveats

As a curated list, the repository does not provide direct functionality. The quality and applicability of the linked resources depend on their original sources. Some links may become outdated over time.