It was simply introduced that Charles Margossian and Lawrence Saul gained the perfect paper award on the 2025 version of the Synthetic Intelligence and Statistics (AISTATS) convention. This information is so contemporary, it’s not but on the awards web page but. They gained for the next paper.
- Charles C. Margossian and Lawrence Ok. Saul. 2025. Variational Inference in Location-Scale Households: Actual Restoration of the Imply and Correlation Matrix. AISTATS.
Right here’s Charles, nattily wearing a blue blazer for the award ceremony:
Photograph credit score: Daniel Lee (the Cornell CS professor, not the Stan developer)
For individuals who won’t know Charles, he was a Ph.D. pupil of Andrew’s, is presently a postdoc at Flatiron Institute, and as he posted right here just lately, is on his approach to begin a school place within the statistics division on the College of British Columbia (Vancouver). I can say from a entrance row seat that Lawrence is basically actually good at teaching writing and presenting and Charles is a superb presenter (although the latter wouldn’t assist with the perfect paper award—they knew about it earlier than the presentation so we received to see the apply speak). In the event you get an opportunity to take considered one of Lawrence’s scientific writing tutorials, take it—it’ll for positive offer you a leg-up in a greatest paper competitors.
It’s a small world in Bayesian-flavored AI and stats. I imagine that’s Stephan Mandt on the fitting of the picture—he’s the final program chair for AISTATS this yr. Stephan was a postdoc with Dave Blei at Columbia whereas I used to be there and labored with Matt Hoffman at Google. Charles has additionally printed papers on VI with Dave Blei. Daniel Lee (who took the picture above) was at Bell Labs with me within the late Nineteen Nineties. I generally bumped into Lawrence on the prepare again then going to AT&T Labs.
