iiinova — where your digital self keeps evolving

Kejun (Albert) Ying

@kejunying.bsky.social

F99K00 Postdoc fellow Wyss-Coray @Stanford & Baker lab @UWproteindesign | Prev. Bio PhD & CS MS @Harvard | BoA Consortium core | Solving aging to solve universe

Posts Replies Media

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

10/10 Thank you to all our collaborators, including @jameszou.bsky.social and the entire HealthFlux team! #AIforScience #AIforAging #WorldModels

000

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

9/10 There is so much more I want to explore with this approach. Thank you to Zirui Wang, Tony Wyss-Coray and the whole team! Tony and I are co-corresponding authors. Paper: doi.org/10.64898/202...

doi.org

100

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

8/10 HealthFlux supports virtual clinical trials: starting from the same health state, we simulate futures with and without a medication to explore changes in physiology and disease risk. A way to explore drug repurposing and prevention before designing clinical studies.

100

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

7/10 The same health state also supported predictions of future laboratory measurements. We want to connect changes in a person’s biology with the diseases that may develop later.

100

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

6/10 Combining measurements helps us see risks that individual inputs miss. In our hypertension analysis, HealthFlux identified future cases missed by models using a single modality.

100

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

5/10 The model could predict diseases excluded from its training. This is one of the findings I’m most excited about: a shared health state can support predictions for conditions it was never trained to predict.

100

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

4/10 HealthFlux achieved the best overall disease-prediction performance among the models we compared. It also kept predicting future risk when we stopped supplying new observations.

100

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

3/10 HealthFlux is a pan-modal world model: 5,647 features across 11 data domains, from clinical records to molecular measurements and imaging. It learns a shared health state that evolves over time and updates as new observations arrive.

100

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

2/10 Disease develops over years, across many parts of the body. We want models that can follow that process and help us understand how someone’s health might change next.

100

Kejun (Albert) Ying @kejunying.bsky.social · 28/09/2026

1/10 Can we build a world model of human health? I'm thrilled to announce my first senior author paper: HealthFlux! Simulating how health evolves over a lifetime to find ways to prevent disease. doi.org/10.64898/202...

110