News
What the lab is up to.
Papers, talks, and workshops from the lab. Longer write-ups live on brianjalaian.com.
- Oct 2026
Co-chairing the LEAF workshop at IEEE MILCOM 2026
Brian Jalaian co-chairs LEAF (Lightweight Edge-Adaptive Foundation Models), Workshop 4 at IEEE MILCOM 2026 near Washington, DC, with Venkat Dasari and Peng Wang of DEVCOM ARL. The lab presents two posters: Measure, Then Compress and GOE.
- Sep 2026
Accept Rules accepted to the NeurIPS 2026 AI for Science Workshop
“Your Research Loop Is Measuring Noise: Accept Rules for Autonomous Experimentation” will be presented as a poster in Sydney, Australia, on December 12.
- Sep 2026
CUSP preprint on arXiv
Johnny Yu and David Garcia released CUSP, which splits a multi-agent vision-language system’s uncertainty into each agent’s own doubt and the disagreement between agents.
- Jun 2026
Huynh Tran gives a lightning talk at the Mayo Clinic AI Research Summit
Huynh presented uncertainty-aware detection of rare leukemia cells in flow cytometry, our collaboration with Mayo Clinic hematopathology, at the 2026 AI Research Summit in Rochester, Minnesota.
- Apr 2026
Johnny Yu presents SCoOP at ICLR 2026
SCoOP, a training-free way to measure how uncertain a team of vision-language models is, appeared as a poster at the ICLR 2026 workshop on Agentic AI in the Wild.
- Oct 2025
Two student-led papers at IEEE MILCOM 2025
Jake Sander presented fine-tuning vs. distillation for compressing LLMs onto edge hardware, and Huynh Tran showed that neurosymbolic transfer learning improves network intrusion detection.