I’m excited to share that our manuscript was accepted to Neurosurgery (@NeurosurgeryCNS)! In our latest work, we use data collected from smartphone-based sensors to evaluate mobility trends among patients with Glioblastoma (an aggressive form of brain cancer) after receiving a

RT @JKJones_: 1/ The US used to dominate the critical minerals space. But we outsourced all of it to China. Tax credit, executive orders a….

RT @JeffDean: My longtime collaborator Dave Patterson (long-time faculty at @UCBerkeley, @TheOfficialACM Turing Award winner, and fellow @L….

Journal Link: Paper Link: Code:

Many thanks to my PhD advisors @jponnela and Junwei Lu (Harvard University), and collaborator Timothy R. Smith (Brigham and Women's Hospital).

We applied KSH-LSPI to a digital phenotyping study of 67 spine surgery patients enrolled at Brigham and Women's Hospital. Our method suggests that early mobilization strategies may enhance long-term surgical outcomes, affirming established postoperative rehabilitation guidelines.

(3) We enable local approximation through kernel-weighted estimation that adapts to different regions of the feature space

(2) We capture complex, nonlinear relationships without restrictive parametric assumptions—allowing us to explore interactions between continuous state features and actions.

(1) We first decompose value functions into sparse, additive components using adaptive regularization via group Lasso penalties for automatic feature selection.

What makes our approach unique? Our KSH-LSPI (Kernel Sieve Hybrid - Least Squares Policy Iteration) method bridges the gap between flexible machine learning and clinical interpretability

🎉 Excited to share that our paper was published in The Annals of Applied Statistics!. "Nonparametric Additive Value Functions: Interpretable Reinforcement Learning with an Application to Surgical Recovery"

RT @MichaelD1729: Consider applying to work with me through @MATSprogram, to help build our understanding of building AI systems towards un….

RT @lhan320: 🚀 How do you transport causal effects w/ survival outcomes? We tackle censoring, discr/cont time & distrib shift in our new pr….

RT @JeffDean: Introducing Gemini 2.0 Flash Thinking, an experimental model that explicitly shows its thoughts. Built on 2.0 Flash’s speed….

RT @chipro: It’s done! 150,000 words, 200+ illustrations, 250 footnotes, and over 1200 reference links. My editor just told me the manuscr….

RT @davidstutz92: If you are at #NeurIPS2024 this week, do check out our work on conformalized credal sets - an efficient way to construct….

Excited to attend NeurIPS this year! If you're passionate about AI in healthcare and interested in building personal healthcare agents for chronic disease management (with a focus on oncology), let's connect! We're currently hiring ML Engineers and a designer.

RT @IlanPrice: GenCast is out in @Nature! It's the first high res ML ensemble weather forecast which outperformed the operational state of….

RT @jparkerholder: Introducing 🧞Genie 2 🧞 - our most capable large-scale foundation world model, which can generate a diverse array of cons….

RT @ChrisGPotts: This talk is now up on YouTube and includes wide-ranging and interesting discussion at the end:

RT @JKJones_: 1/🧵 The race for the best foundation model launched with ChatGPT 3’s release in 2022, and it’s been nonstop ever since. 🚀 But….