Neural Entropy

2024/09/05 by Akhil Premkumar, Premkumar, Akhil · 1 voice · 4 citations
Medicine · Computer Science · #Advanced Neuroimaging Techniques and Applications #Generative Adversarial Networks and Image Synthesis #Stochastic Gradient Optimization Techniques

paper · pdf · doi:10.48550/arxiv.2409.03817

Abstract

We explore the connection between deep learning and information theory through the paradigm of diffusion models. A diffusion model converts noise into structured data by reinstating, imperfectly, information that is erased when data was diffused to noise. This information is stored in a neural network during training. We quantify this information by introducing a measure called neural entropy, which is related to the total entropy produced by diffusion. Neural entropy is a function of not just the data distribution, but also the diffusive process itself. Measurements of neural entropy on a few simple image diffusion models reveal that they are extremely efficient at compressing large ensembles of structured data.

Cited by

• Stochastic Thermodynamics for Autoregressive Generative Models: A Non-Markovian Perspective
• The Information Dynamics of Generative Diffusion
• Entropic Time Schedulers for Generative Diffusion Models
• Mutual Information Estimation via Score-to-Fisher Bridge for Nonlinear Gaussian Noise Channels

Discussions

• Tomorrow (Monday) Akhil Premkumar will present his "Neural Entropy" arxiv.org/abs/2409.03817 Join us on zoom at 9am PT / 12pm ET / 6pm CET: portal.valencelabs.com/logg [bsky, 18 points, 0 comments]

Related

• TENDE: Transfer Entropy Neural Diffusion Estimation
• Entropy-based Guidance of Deep Neural Networks for Accelerated Convergence and Improved Performance
• An Introduction to Neural Data Compression
• NoisyNN: Exploring the Impact of Information Entropy Change in Learning Systems
• Modeling Information Flow Through Deep Neural Networks