This repository is a curated list of academic papers related to the Information Bottleneck (IB) principle, a framework that explains deep learning phenomena through the lens of information theory. It serves researchers and practitioners interested in understanding and applying IB to areas like neural network analysis, representation learning, and reinforcement learning.

How It Works

The repository organizes papers into thematic categories, including classics, reviews, theoretical foundations, models, applications (general and RL), and methods for mutual information estimation. This structure allows users to navigate the breadth of IB research, from foundational concepts to cutting-edge applications and the critical challenge of estimating mutual information.

Quick Start & Requirements

This is a curated list of papers, not a software package. No installation or execution is required.

Highlighted Details

• Comprehensive categorization of IB research, including seminal works and recent advancements.
• Dedicated sections for mutual information estimation techniques, a key challenge in IB.
• Includes links to papers, code, slides, and talks where available.
• Covers applications across various domains, including deep learning, reinforcement learning, and neuroscience.

Maintenance & Community

The repository was last updated in October 2022. Contributions are welcomed via pull requests.

Licensing & Compatibility

This repository contains links to academic papers and does not have its own software license. The licensing of individual papers would need to be checked on their respective publication platforms.

Limitations & Caveats

The list is a snapshot as of October 2022 and may not include the very latest research. The primary focus is on academic literature, not executable code or practical tools.