Photo by Toa Heftiba on Unsplash

Deep Delight

Combines Reinforcement Learning with DLITE (Delight) as a new information measure (loss function) for training and fine-tuning of machine/deep learning models.

Deep Learning Information Theory Machine Learning
Weimao Ke
Associate Professor of Information Science

My research connects information retrieval, information theory, distributed and agentic AI, and privacy-preserving local language models.

Next
Previous