The A3D3 Institute is hosting the High-Throughput AI Methods and Infrastructure Workshop.

The purpose of this workshop is to exchange and explore high-throughput AI tools and infrastructure across the diverse research disciplines related to the A3D3 (Accelerated Artificial Intelligence Algorithms for Data-Driven Discovery) Institute with the goal of fostering multi-disciplinary discussion and idea exchange. Participants will discuss the high-throughput AI tools, platforms, and challenges related to their ongoing research and work together to grow and apply their knowledge to a hackathon challenge. This is an opportunity for students, postdocs, and researchers to effectively communicate research, collaborate, and apply AI tools in a hands-on and cross-disciplinary environment.

Learn more and register here.

Participants must pre-register for the workshop. There is no registration fee.

This workshop will take place over a week, starting with an overview on Monday of A3D3 research areas and outline of AI methods and platforms within those research areas (neuroscience, high energy physics, multi-messenger astronomy). Participants will then have the opportunity to present posters on related research during the workshop reception on Monday night, and will present in two groups so they will be able to view and provide feedback on each other’s posters.

The workshop will resume on Wednesday, with the kickoff of a hackathon challenge. This will be a “kaggle-style” AI competition judging the performance and latency of a machine learning model deployed on an FPGA (specific topic to be announced). Cross-disciplinary and multi-experience-level teams will be formed and have the opportunity to collaborate on this task on Wednesday and Thursday before presenting their progress on Friday at the closing of the workshop. Winners will be announced in a remote event at the conclusion of the challenge in the fall. Teams will be encouraged to meet regularly to complete their project post-workshop.

There will also be lab tour visits on Wednesday and Thursday afternoon, respectively.