Drew Prinster

Safe & regulatable AI/ML for health | My job is (mostly) error bars 🫡 (eg, conformal prediction) | CS PhD at Johns Hopkins. Prev at Yale. he/him

[Drew Prinster @DrewPrinster]

Can we ensure AI agents respect our safety constraints, even as they explore & improve?
- Medical LLMs that are helpful, & avoid false claims?
- Bioscience agents that generate effective molecule designs, & ensure they’re safe?
📄🧵w/ @samuel_stanton_ @clara_fannjiang @jiwoncpark @kchonyc @anqi_liu33 @suchisaria

Excited to share “Conformal Policy Control” ⬇️

1/12

[Drew Prinster @DrewPrinster]

@StephenLCasper @MITCSAIL @Harvard @Kennedy_School Congrats! :)

[Drew Prinster @DrewPrinster]

@tifding @stats_stephen Congrats!!

[Samuel Stanton @samuel_stanton_]

accepted to ICML as a spotlight, see you in Seoul!

[Anqi (Angie) Liu @anqi_liu33]

Gina will be attending AISTAT presenting this paper. She is also close to graduating and looking for her next step. Talk to her there!

[Hannah Lawrence @HLawrenceCS]

TL;DR: poster today at 3:15pm, P3-#1109!

Have you ever benchmarked your method on QM9, MD17, OC20, or ModelNet? It turns out that the 3D orientations of point clouds in commonly used datasets are highly non-random. How did we prove this and why should you care? 🧵

[Drew Prinster @DrewPrinster]

@williamjurayj Thanks for the kind words Will! 'Hope and think there's some cool opportunities for extensions and applications in this space :)

[Jess Sorrell @OptimistsInc]

I took ~7 years off during undergrad. Worked at Starbucks, the postal service, a diner. Wasn't until making friends with some CS PhD students at UW Madison, who suggested sitting in on Eric Bach's class on the physics of computation, that I decided to go back (and then get a PhD)

[Samuel Stanton @samuel_stanton_]

You have a safe model you've tested, and you have a new post-trained model. How far can you trust a new model before it becomes unsafe? The answer goes to the heart of statistical decision-making. To be safe, the agent must be self-aware. Read @DrewPrinster's thread for more.

[Drew Prinster @DrewPrinster]

Many many thanks to the absolutely amazing team of collaborators on this project, who I was so privileged and lucky to work with! This couldn’t have happened without numerous and major contributions from @samuel_stanton_ @clara_fannjiang @jiwoncpark @kchonyc @anqi_liu33 @suchisaria

And, sincere thanks to support from Prescient Design @genentech, @HopkinsDSAI, and @MooreFound !

12/12

[Drew Prinster @DrewPrinster]

Can we ensure AI agents respect our safety constraints, even as they explore & improve?
- Medical LLMs that are helpful, & avoid false claims?
- Bioscience agents that generate effective molecule designs, & ensure they’re safe?
📄🧵w/ @samuel_stanton_ @clara_fannjiang @jiwoncpark @kchonyc @anqi_liu33 @suchisaria

Excited to share “Conformal Policy Control” ⬇️

1/12

[Drew Prinster @DrewPrinster]

Safety and exploration, in this view, can complement rather than oppose each other. With the right balance they can encourage progress while avoiding pitfalls. This balance, however, must be determined from what we actually know, not from what we hope. It can be enough, it turns out, to know what is safe, and to know what we want to try next.

11/12