IFML Hosts Fall 2025 AI Research Symposium

The Fall AI Research Symposium, held on November 5, 2025, at the Mulva auditorium on the UT campus, marked five years of groundbreaking work from the NSF National AI Institute for Foundations of Machine Learning (IFML). The event also marked the recent renewal of IFML’s future funding to support breakthrough research in generative AI that is open-sourced, enabling broader adoption and innovation across a wide range of fields. IFML director Adam Klivans welcomed guests with remarks noting key accomplishments of the institute such as its 119 active research projects, more than 100 accepted papers, over 150 IFML seminars, and multimedia coverage of IFML’s work through numerous interviews, TV news stories and podcasts. Over the past 5 years, IFML has had a strong research presence and has acted as a nexus point across various institutions. It has supported a wide variety of researchers and over 40 graduate students, and has formed collaborations with industry partners such as Meta, Amazon, Dell, Google, YouTube, amongst others.

In addition to celebrating another 5-year renewal, the event also convened leading researchers to share their perspectives on the most transformative developments in generative AI. Sanjay Shakkotai, UT Austin Professor Cockrell Family Chair in Engineering #15, and Director of the Center for Generative AI, kicked off the event with a presentation on diffusion models and their applications in generative AI. Visiting keynote speaker, Amin Karbasi, Senior Director of AI Research, Cisco Foundation AI, presented on “Reasoning: A Double-Edged Sword,” and how reasoning is taking AI models to new heights while at the same time exposing them to new risks.

The afternoon’s keynote speaker, Vahab Mirrokni, Google Fellow and VP at Google Research, provided an overview  of how algorithms can address key challenges in Generative AI and Large Language Models (LLMs).

Postdoctoral Research Fellow, Danny Diaz, presented ground breaking research from IFML’s Deep Proteins group. His talk addressed the use of AI in developing protein-based biotechnology, while UT Austin Professor Kristen Grauman shared highlights  from her transformative research on multimodal activity understanding and the use of video understanding. 

The symposium culminated in a thought-provoking panel discussion on controversial and emerging topics in AI with Adam Klivans, Sujay Sanghavi, Danny Diaz, Atlas Wang and Vahab Mirrokni.