Michigan AI

Eleven papers from University of Michigan College of Engineering CSE of Michigan CSE are being presented at ! 🚀

These contributions span multimodal agents, language models, AI safety, deep learning theory, robotic memory, and more, highlighting the breadth of Michigan's machine learning research.

https://cse.engin.umich.edu/stories/eleven-papers-by-cse-researchers-at-icml-2026

Start your week with a deep dive into decades of research on creating agents with the full range of human cognitive abilities.

Don't miss this interview by Tom Mitchell w/ Univ. of Michigan CSE Prof. Emeritus John Laird:⬇️
🍎 https://podcasts.apple.com/us/podcast/ai-agents-to-model-human-cognition-with-john-laird/id1879433175?i=1000767180690

📺 https://www.youtube.com/watch?v=YP2YbF93-1Y

AI Agents to Model Human Cognition with John Laird Podcast Episode · Machine Learning: How Did We Get Here? · May 11 · 33m

⚡ Does language influence AI's energy use?

A new study, The Language–Energy Divide, finds that it does.

Analyzing LLM inference in 122 languages, it identifies a systemic disparity in energy consumption, highlighting an overlooked challenge for building more efficient and equitable AI systems.

Work by Naihao Deng, Alissa Shen, Yiming Feng, Joan Nwatu, Jae-Won Chung, Mosharaf Chowdhury, Yulong Chen, Rada Mihalcea.

📄 Paper: https: //arxiv.org/abs/2606.21869
🌐 Project page: https://t.co/TgPWzwZePl