Johns Hopkins University Computer Science

09/04/2026

ICYMI: CS alumnus Jieneng Chen has been recognized by Scientific American as a Young American Scientist for his work in computer vision.

The inaugural Young American Scientists class of 28 early-career researchers spans a broad array of fields, including healthcare, botany, and astrophysics.

The inaugural Young American Scientists class of 28 early-career researchers spans a broad array of fields, including healthcare, botany, and astrophysics.

09/03/2026

Our computer scientists found that large language models often continue to rely on information learned during training—even when they’ve been instructed to use new information. 🤔

A Johns Hopkins study reveals how large language models fail to ignore information learned during training.

09/03/2026

Congratulations, Prof. Zhou! 🌟

Zhou’s research focuses on medical computer vision, language, and graphics for early cancer detection and diagnosis. He is best known for developing UNet++, a widely adopted segmentation architecture ranked among the most popular articles in IEEE Transactions on Medical Imaging.

09/02/2026

Interested in ? Learn about the work our researchers will be presenting next week at in Sweden! 🇸🇪 (Post 2 of 2)

In “MemoBench: Benchmarking World Modeling in Dynamically Changing Environments,” Wufei Ma, Alan Yuille, and collaborators from Harvard University, the Massachusetts Institute of Technology (MIT), Boston University, Google, and Carnegie Mellon University introduce a diagnostic benchmark built around the disappear-and-reappear 🪄 paradigm in dynamically changing environments: https://arxiv.org/abs/2606.27537

“PhyGDPO: Physics-Aware Groupwise Direct Preference Optimization for Physically Consistent Text-to-Video Generation” by Yuanhao Cai, Alan Yuille, and collaborators at Meta and The Chinese University of Hong Kong 香港中文大學 - CUHK formulates a framework that uses real-world video 📹 to guarantee correct physics learning: https://arxiv.org/abs/2512.24551

Chen Wei, Alan Yuille, and researchers from The Hong Kong University of Science and Technology - HKUST and Fudan University propose an efficient reward modeling framework that eliminates manual preference annotation and explicit quality dimension engineering in “Fake it Till You Make it: Reward Modeling as Discriminative Prediction”: https://arxiv.org/abs/2506.13846

In “FaceMoE: Mixture of Experts for Low-Resolution Face Recognition,” Kartik Narayan and Vishal M. Patel propose an effective adaptation of Mixture of Experts transformer architecture for low-resolution face 😀 recognition: https://arxiv.org/abs/2606.32040

“Silhouette-Based Gait Foundation Model” by Dingqiang Ye, Kartik Narayan, Vishal M. Patel, and colleagues at Shenzhen University introduces the first scalable, self-supervised pretraining framework for gait 🚶 understanding: https://arxiv.org/abs/2512.00691

Deming Li, Abhay Yadav, Rama Chellappa, Anand Bhattad, and Cheng Peng present a framework that enforces cross-view consistency during diffusion-based refinement of reconstructed scenes in “SyncFix: Fixing 3D Reconstructions via Multi-View Synchronization”: https://arxiv.org/abs/2604.11797

and Guoyizhe Wei, Feng Wang, Alan Yuille, and Rama Chellappa will present “Multi-Head Normalization for Wide Vision Transformers.”

09/02/2026

Interested in ? Learn about the work our researchers will be presenting next week at in Sweden! 🇸🇪 (Post 1 of 2)

In “Imaginative Perception Tokens Enhance Spatial Reasoning in Multimodal Language Models,” Jaemin Cho and collaborators at the University of Washington, the Allen Institute for AI (AI2), Microsoft, and OpenAI introduce intermediate perceptual representations that externalize what a vision-language model would perceive under alternative spatial configurations: https://arxiv.org/abs/2606.03988

“AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories” by Jaemin Cho and researchers from The University of North Carolina at Chapel Hill (UNC-CH) and Nanyang Technological University, Singapore introduces a memory-augmented video generation framework that can reconcile cross-view inconsistencies: https://arxiv.org/abs/2602.14941

Jaemin Cho and colleagues from UNC-CH, New York University, Meta, and AI2 present a systematic study of visual co-denoising in a unified JiT-based framework in “V-Co: A Closer Look at Visual Representation Alignment via Co-Denoising”: https://arxiv.org/abs/2603.16792

In “Physics Question Scene Graph: Fine-Grained Evaluation of Physical Plausibility in Text-to-Video Generation” Jaemin Cho, Elias Stengel-Eskin, Engr ’23 (PhD), and UNC-CH and AI2 researchers introduce a hierarchical question-based evaluation pipeline for video generation models: https://arxiv.org/abs/2606.25306

“Shared LoRA Subspaces for Almost Strict Continual Learning” by Prakhar Kaushik, Ankit Vaidya, Shravan Chaudhari, Rama Chellappa, and Alan Yuille proposes a novel approach to parameter-efficient continual fine-tuning: https://arxiv.org/abs/2602.06043

Jonathan Lee, Xingrui Wang, Jiawei Peng, Luoxin Ye, Zehan Zheng, Tiezheng Zhang, Tao Wang, Wufei Ma, Siyi Chen, Yu-Cheng Chou, Prakhar Kaushik, and Alan Yuille propose a structured process of scene understanding in “Perceptual Taxonomy: Evaluating and Guiding Hierarchical Scene Reasoning in Vision-Language Models”: https://arxiv.org/abs/2511.19526

09/02/2026

Behind the scenes with Suchi Saria and Jake Tapper! Watch the CNN interview on her AI-powered sepsis detection tool here: https://www.cnn.com/2026/08/26/health/video/sepsis-monitor-detection-artificial-intelligence-fda-lead-jake-tapper

09/01/2026

Don’t miss this year’s ACM Lecture in Memory of Nathan Krasnopoler with Adobe’s Titus Winters! Learn more here: https://www.cs.jhu.edu/event/acm-lecture-in-memory-of-nathan-krasnopoler-cs-education-post-ai/

09/01/2026

Congratulations to our BDP Summer Fellows on completing their summer research!

The Bloomberg Distinguished Professorships Summer Fellowship Program gives Johns Hopkins undergraduates the opportunity to work and conduct research under the mentorship of some of the world’s preeminent scholars.

The Bloomberg Distinguished Professorships summer fellowship program gives Johns Hopkins undergraduates the opportunity to work and conduct research under the mentorship of some of the world’s preeminent scholars.

08/31/2026

Researchers from Prof. Chien-Ming Huang’s Intuitive Computing Laboratory shared their storytelling robot, ELLA, with young learners at Henderson-Hopkins this summer! 🤖

The students listened to interactive stories, answered questions, and even tested the robot with their own math questions! ➗

Learn more about ELLA here: https://dl.acm.org/doi/epdf/10.1145/3773077.3806113

08/31/2026

Featuring Johns Hopkins University’s Suchi Saria:

10 most influential Indians in Time100 AI 2026: Meet the Indian tech leaders shaping AI’s future Read here 👉 https://indianexpress.com/article/trending/top-10-listing/10-most-influential-indians-in-time100-ai-2026-meet-the-indian-tech-leaders-shaping-ais-future-10854327/?utm_term=Autofeed&utm_medium=echobox&utm_channel=echoboxfb&utm_source=Facebook =1787988708