热点大瓜

The SUNY AI Symposium aims to generate new ideas and partnerships in the rapidly evolving field of artificial intelligence, and no time was wasted doing just that. 

The two-day event opened Tuesday morning at UB with keynote speaker Damien Fair discussing how computing power is helping his team at the University of Minnesota better understand brain function.

Given that UB is currently using AI to assist children with speech and language disorders, the symposium’s co-host, Venu Govindaraju, vice president for research and economic development, told Fair that he sees great potential for collaboration.

The second annual symposium, which concludes today, brings together researchers from across the SUNY system and other universities for discussions on how to harness the power of AI for the betterment of society. 

President Satish K. Tripathi welcomed those gathered in Lippes Concert Hall in Slee Hall for the symposium’s opening.

“We greatly look forward to learning more about your AI work and the discoveries it is yielding for the public good,” he said. “In short, that is what will be exploring over the next two days: how we are harnessing the power of AI to contribute meaningfully to every sector of society.”

Tripathi also highlighted the work happening at UB, where more than 200 faculty members are conducting AI studies to improve childhood literacy, predict disease progression, ensure information integrity and mitigate climate change, among many other projects.

“And now, as the home of Empire AI, we will be able to tap the full power of artificial intelligence to solve even more of the complex issues facing our state, nation and world,” Tripathi said.

An update on Empire AI was provided by SUNY Research Foundation President Melur "Ram" Ramasubramanian. He described some of the research consortium’s projects that are already underway thanks to supercomputing power housed at UB.

Empire AI’s Alpha system, which went online this past October at the Downtown Campus, is already among the top 250 most powerful computers in the world. It will be complemented this fall by the addition of the Beta system, which will be up to seven times faster.

“So it is time for us to imagine problems that have larger and larger datasets and really look at a scale that we would not have planned to look at [previously] due to lack of resources,” Ramasubramanian said.

Both Alpha and Beta will eventually be moved to the new supercomputing center to be built on the North Campus. The facility’s conceptual design will be completed this summer and include capabilities for adding further computing power throughout its operational life, Ramasubramanian said.

There were several other updates about SUNY’s push into AI, including the development of SUNY GPT. This new SUNY-based, generative AI chatbot program can be tailored for coursework, research and student projects. 

“Students will be able to develop projects, learn from faculty and peers, and interact with customized AI-based tutors who can help them anywhere, anytime, all while maintaining strong protections and standards to ensure academic access,” said SUNY Chancellor John B. King Jr. in a video message. 

Shadi Sandvik, senior vice chancellor for research, innovation and economic development and the symposium’s co-host, spoke about the SUNY INSPIRE Center.

“This center of centers will serve as a hub for scaling AI education research and scholarship, with a focus on attracting funding and advancing AI for public good,” she said. “This spring, we are formalizing the center with the launch of its governance and science and technology advisory groups.”

Tuesday’s event also included eight short presentations from AI experts across the SUNY system, including UB faculty members X. Christine Wang (AI and education), Siwei Lyu (AI and security) and Sophie Nowicki (AI and climate), as well as a discussion involving AI and the law with UB’s Matthew Pelkey and SUNY law colleagues.

Fair, the event’s keynote speaker, is a professor at the University of Minnesota Medical School and director of the Masonic Institute for the Developing Brain, who uses the latest brain imaging to create personalized maps of patients’ brains. These maps can then be used to surgically implant electrodes into specific areas to treat issues like depression and chronic pain, a treatment called stimulation therapy. 

“Neuroscience, computing, engineering, modeling and industry financing all crashed together in one spot in this moment in time, and that’s allowing us to do things that we thought were unimaginable, even just a few years ago,” Fair said.

AI is used to fine-tune the stimulation settings of the electrodes. Patients and doctors first work to find the best settings for them based on their mood and sleep. 

“We then put that into the AI model and it will generate new selections around what was best, and we start to zero in over time until we get the best setting for the patient,” Fair said.

He played a video of one of his patients trying out different settings following surgery. The patient is a man in his 40s with severe depression that prevented him from experiencing the full spectrum of emotions.

In the video, he is overcome with emotion and says, “It feels good. … I don’t know what to do with it.”

When the team asks if they should turn the stimulation down, he quickly says, “No. Let’s stay here for a second.”

When asked to describe what he is feeling, he says, “It’s a joy.”