Johns Hopkins Data Science and AI Institute

Tonight: James Bellingham, Bloomberg Distinguished Professor and Director of Research for Safety and Assurance at the Data Science and AI Institute, will speak on “How Are Marine Robots Shaping Our Future?”

Congratulations to @ananya-a-joshi.bsky.social on winning a
@aaai.org Deployed AI System Award for this work, which proposes a system that reduces alert fatigue in big-data decision making.

Researchers use AI tools to model, improve wildfire evacuation Scientists at Hopkins, University of Florida simulate and predict human behavior during wildfire evacuation, allowing for improved planning and safety

New AI research is helping predict how people behave during wildfire evacuation, so emergency planners can reduce traffic jams and save lives. This federally funded research is helping to expand AI models grounded in human behavior to improve disaster response. https://bit.ly/4rzN1Fu

AI to enhance sports The Johns Hopkins Sports Analytics Research Group’s innovations were among those featured at the inaugural World Changing Ideas Summit.

As featured at the inaugural @jhu.edu / @fastcompany.com World Changing Ideas summit, the Johns Hopkins Sports Analytics Research Group is leading real-world sports applications of #AI.

Headshots of Anqi “Angie“ Liu and Tom Lippincott.

Learn how @aliu33.bsky.social and @tom-lippincott.bsky.social are using their @hopkinsdsai.bsky.social Seed Grant Awards to address a broad range of societal and humanitarian challenges through the use of data science and #AI: www.cs.jhu.edu/news/dsai-20...

As #AI transforms nearly every sector, two @jhu.edu experts—including JHU CS’ Suchi Saria—outline guiding principles for its use in medical care so that it can truly benefit physicians and patients. Read the commentary on @natmed.nature.com:

AI predictions for 2026 Johns Hopkins artificial intelligence expert outlines what policymakers should expect in AI over the next year.

What to expect in AI in 2026? @mdredze.bsky.social, director of the Data Science and AI Institute, discusses what’s next for AI and what policymakers should keep in mind over the next year.
washingtondc.jhu.edu/news/ai-in-2...

Join us in advancing data science and AI research! The Johns Hopkins Data Science and AI Institute Postdoctoral Fellowship Program is now accepting applications for the 2026–2027 academic year. Apply now! Deadline: Jan 23, 2026. Details and apply: apply.interfolio.com/179059

For 150 years, Johns Hopkins University has redefined what is possible through research, education, and discovery. As we enter our Sesquicentennial year, it's time to celebrate the relentless pursuit of knowledge and the people behind it. 150.jhu.edu

Join us in advancing data science and AI research! The Johns Hopkins Data Science and AI Institute Postdoctoral Fellowship Program is now accepting applications for the 2026–2027 academic year. Apply now! Deadline: Jan 23, 2026. Details and apply: apply.interfolio.com/179059

Check out the Fall 2025 Hopkins Engineering magazine featuring the latest updates on the Johns Hopkins Data Science and AI Institute!

Alexis Battle speaks into a microphone.

Attendees sit around a round table with a number 6 card on it and discuss. "Pitch a Problem" and various topics are projected onto a screen behind them.

Pedro R.A.S. Bassi presents a poster to an onlooker.

An audience member asks a panel (comprising Paul Yi, Andrew Menard, and Haris Sair) a question.

Last week at the annual Johns Hopkins Research Symposium on Engineering in Healthcare, we learned about developing and deploying #AI tools for clinical practice, discussed emerging challenges, built connections, and fostered new collaborations. Stay tuned for session recordings!

Congratulations to @jhu.edu researchers who have been selected as a 2025 Humanities and AI Virtual Institute awardee by @schmidtsciences.bsky.social ! JHU collaborators will develop an #AI toolkit to help scholars analyze hierarchical patterns in language and music. Learn more below:

Johns Hopkins Joins DOE’s New AI-Powered Scientific Discovery Institute - Johns Hopkins - Ralph O’Connor Sustainable Energy Institute Johns Hopkins is part of the new multi-institution Learning Accelerated Domain Sciences (LEADS) Institute, a Department of Energy-funded collaborative initiative focused...

JHU is part of the new Learning Accelerated Domain Sciences (LEADS) Institute! DOE-funded, LEADS focuses on reshaping how AI supports scientific discovery. ROSEI's @drgona.bsky.social and Enrique Mallada, as well as Mahyar Fazlyab are involved. #HopkinsEnergy energyinstitute.jhu.edu/johns-hopkin...

Hold on to your PanTS—there’s a new pancreatic cancer detection dataset in town Developed by a Johns Hopkins-led research team, the Pancreatic Tumor Segmentation Dataset may be the key for training AI models to detect pancreatic cancer early enough to make a difference in patient...

At @neuripsconf.bsky.social this year, a @jhu.edu-led research team presented the Pancreatic Tumor Segmentation Dataset—which may be the key for training #AI models to detect #pancreaticcancer early enough to make a difference in patients’ survival.

Johns Hopkins APL and Microsoft’s AI Agent Orchestrates Robotic Teams APL and Microsoft recently demonstrated an AI agent that can coordinate heterogeneous robot teams using large language models. The September demo highlighted how APL’s expertise in autonomy software a...

APL and @microsoft.com recently showcased MAESTRO, an agentic AI planner that coordinates diverse robot teams. The demo highlighted how large language models can help robotic systems collaborate in complex environments. https://jhuapl.link/976

AttentiveGRUAE: An Attention-Based GRU Autoencoder for Temporal Clustering and Behavioral Characterization of Depression from Wearable Data In this study, we present AttentiveGRUAE, a novel attention-based gated recurrent unit (GRU) autoencoder designed for temporal clustering and prediction of outcome from longitudinal wearable data. Our...

AttentiveGRUAE: An Attention-Based GRU Autoencoder for Temporal Clustering and Behavioral Characterization of Depression from Wearable Data Nidhi Soley, Vishal M. Patel, and Casey O. Taylor
arxiv.org/abs/2510.02558

Cover of Johns Hopkins Magazine featuring a person wearing a black shirt, looking thoughtfully to the side, with text that highlights a feature story on their path to AI and health care research.

Johns Hopkins researcher Suchi Saria is leading the way in using AI to make health care safer and smarter. From predicting sepsis hours earlier to scaling real-time clinical tools, her work is transforming how hospitals save lives. Read more in the latest issue of Johns Hopkins

The Platonic Universe: Do Foundation Models See the Same Sky?
Kshitij Duraphe, Michael J. Smith, Shashwat Sourav, John F. Wu
www.arxiv.org/abs/2509.19453

The Explore-Exploit Tradeoff Redefined: Balancing Regret and Treatment Effects in Contextual Bandits
Alec Koppel, Sujay Bhatt, Sihan Zeng, Sumitra Ganesh
constrained-opt-ml.github.io/papers/

Solving Bi-Level Reinforcement Learning: A Regularized Actor-Critic Algorithm with Finite-Sample Analysis
Sihan Zeng, Sujay Bhatt, Sumitra Ganesh, Alec Koppel
arlet-workshop.github.io/neurips2025/...

Re-envisioning Euclid Galaxy Morphology: Identifying and Interpreting Features with Sparse Autoencoders
John F. Wu, Michael Walmsley
www.arxiv.org/abs/2510.23749

Migration as a probe: A generalizable benchmark framework for specialist vs. generalist machine-learned force fields
Yi Cao, Paulette Clancy
sites.google.com/view/ai4mat/...

Learning to Optimize for Mixed-Integer Non-linear Programming with Feasibility Guarantees
Bo Tang, Elias B. Khalil, Ján Drgoňa
arxiv.org/abs/2410.11061

Coupling Language Models with Physics-based Simulation for Synthesis of Inorganic Materials
Edward W. Staley, Tom Arbaugh, Michael Pekala, Alex New, Christopher D. Stiles, Nam Q. Le, Gregory Bassen, Wyatt Bunstine, Tyrel McQueen
openreview.net/forum?id=cty...

Causal Masking on Spatial Data: An Information-Theoretic Case for Learning Spatial Datasets with Unimodal Language Models
Jared Junkin, Samuel Nathanson
arxiv.org/abs/2510.27009

Whitened Score Diffusion: A Structured Prior for Imaging Inverse Problems
Jeffrey Alido, Tongyu Li, Yu Sun, and Lei Tian
neurips.cc/virtual/2025...

When Does Curriculum Learning Help? A Theoretical Perspective
Kaibo Zhang, Yunjuan Wang, Raman Arora
neurips.cc/virtual/2025...

Visual Jenga: Discovering Object Dependencies via Counterfactual Inpainting
Anand Bhattad, Konpat Preechakul, Alexei A. Efros
neurips.cc/virtual/2025...

Vision Language Vision Auto Encoder: Scalable Knowledge Distillation from Diffusion Models
Tiezheng Zhang, Yitong Li, Yu-Cheng Chou, Jieneng Chen, Alan Yuille, Chen Wei, Junfei Xiao
neurips.cc/virtual/2025...