Michigan AI

Nominations are now open for Rising Stars in AI at the 9th Annual Michigan AI Symposium on OCTOBER 20, 2026!

If you advise/mentor postdoc or a PhD student (4th+ year) who's aiming at an academic career, please nominate them using the form below. Selected nominees are invited to give a research talk and present a poster at the symposium in front of a mix of academic and industry researchers. Nominees from any area of are encouraged. Nominations close AUGUST 31.

➡️https://tinyurl.com/44csd8jx

Interested in participating in the Rackham Doctoral Intern Fellowship Program? Winter deadlines are coming up quickly:

Track 1: Internships with host organizations, identified by Rackham.
➡️ Winter 2027 Deadline: October 16

Track 2: Internships that Rackham students self-identify and secure on their own.
➡️ Winter 2027: October 23

Learn more: myumi.ch/gR9PG

Less data, smarter training. 🧠

This paper introduces a skill-aware data selection and fine-tuning framework that transfers reasoning abilities from large models to smaller ones using just 1,000 carefully selected examples.

Check it out: https://aclanthology.org/2026.acl-short.49/

Smarter fusion for graph learning !💡

GLANCE brings adaptive, node-aware GNN+LLM fusion, improving results across benchmarks while keeping costs low.
Work by Donald Loveland, Yao-An Yang & Danai Koutra:

https://arxiv.org/abs/2510.10849

📣 Call for Poster and Demo Contributions for Michigan AI Symposium 2026!

This annual symposium brings together leading researchers, students, industry professionals, and thought leaders.

Submit by: SEPTEMBER 21 (registration required)
➡️https://lnkd.in/dM7xWeP7

🤖 What does it take to build a robot from scratch?

At our next Friday Night AI, we'll explore the engineering behind intelligent robots with Ann Arbor Public Schools students. Join us and learn what it takes to build robots that can navigate the real world.
Moderator: Prof. Rada Mihalcea, Michigan AI LAB Director

SEPTEMBER 11 | 6:30pm | AADL downtown (4th floor)
https://ai.engin.umich.edu/events/friday-night-ai/september-11-2026/

How did today's advances in natural language processing become possible?

This article recently published by "Communications of the ACM" explores the pivotal role of support in driving decades of foundational research, training future leaders, and enabling innovations with broad societal impact. This article benefited from the contributions of many individuals, including Michigan AI's Director Prof. Rada Mihalcea.

https://cacm.acm.org/federal-funding-of-academic-research/natural-language-processing-how-nsf-supports-visionary-new-approaches-to-nlp/

are changing how we decide what to trust.

This video from our Friday Night AI “ , , and the Future of Trust” features expert perspectives from Univ. of Michigan Profs. Cliff Lampe & Khalid Malik.

Moderator: Michigan AI Director Rada Mihalcea

Friday Night AI | Deepfakes, AI, and the Future of Trust Advances in artificial intelligence have made it easier than ever to generate convincing synthetic audio, images, and video, blurring the line between what i...

Chimaobi Okite, Anika Misra, Joyce Chai, and Rada Mihalcea introduce a benchmark on lifelong personalization ➡️ LUCid: Redefining Relevance for Lifelong Personalization

LUCid includes 1,936 realistic queries paired with interaction histories of up to 500 sessions, designed to measure situational user-centric relevance in personalization.

✅ Paper: https://arxiv.org/abs/2604.26996

✅ Code: https://t.co/hMc54vTicR

✅ Data: https://t.co/8RFxVRhTHy