The two-day BLASS workshop focused on key advancements in AI for science and engineering. This year’s theme was spatiotemporal data challenges in various fields, such as climate science, seismology, and fluid mechanics.

“The Bay Area is a global hub for AI and scientific research, with leading institutions like UC Berkeley, Stanford, and national labs such as Berkeley Lab, Lawrence Livermore, and SLAC working alongside tech giants like Google, Meta, and others. With BLASS, we aimed to bring this expertise together in an intimate setting to spark conversation and foster collaborations that will tackle the complex scientific challenges shaping the future of research and innovation,” said Ben Erichson, BLASS Co-Organizer and a researcher in the Machine Learning and Analytics Group in Berkeley Lab’s Scientific Data Division.

According to Erichson, he and BLASS co-organizer Wahid Bhimji, who leads the National Energy Research Scientific Computing Center’s (NERSC’s) Data and AI Services Group, have both attended and contributed to numerous AI workshops and meetings. They noticed strong enthusiasm for collaboration between the machine learning and scientific communities but saw the need for a more research-oriented event to foster meaningful collaboration. That’s when the idea for BLASS was born.

In addition to program talks, the summit also included a poster session reception hosted in collaboration with the International Computer Science Institute (ICSI). Bhimji notes that this aspect of BLASS offered an opportunity for attendees to mingle and learn from each other.

“The excitement around this topic was palpable throughout the summit—you could see lively conversations and people lining up to ask our speakers questions. Although the poster started at 4 p.m., many people stayed well into the evening to talk to each other and learn about the work that was going on in this space,” said Erichson.

Bhimji sees BLASS as part of a broader ecosystem of complementary events designed to help the scientific research community harness emerging AI technologies. For developing community expertise in AI for science, Berkeley Lab plans to host an AI for Science Bootcamp next June that will be a follow-up of previous lab-hosted deep learning for science summer schools. BLASS, which Bhimji and Erichson plan to make an annual event, will foster collaborations between experienced researchers and industry partners.

“In discussions about national lab and academic collaborations with industry partners, many have raised concerns about data privacy, cybersecurity, and a host of other issues. My personal view is that these concerns shouldn’t deter us from collaborating. By having an organization like Berkeley Lab serve as the nexus connecting these communities through events like BLASS, we can ensure we’re central to addressing those concerns,” said Bhimji.

In addition to Bhimji and Erichson, the other BLASS co-organizers were Berkeley Lab Applied Math and Computational Research Division (AMCR) Research Scientist Zhi Jackie Yao, NERSC Machine Learning Engineer Steven Farrell, and NERSC Postdoc Vinicius Mikuni.