Johns Hopkins Engineering
@HopkinsEngineer
Followers
18K
Following
8K
Media
2K
Statuses
23K
Research Saves Lives. The @JohnsHopkins Whiting School of Engineering is a global leader in engineering education and research. RTs are not endorsements.
Baltimore, MD
Joined March 2009
We are transforming society—integrating research, education, and translation to give students opportunities to solve problems and make an impact on the world. #HopkinsEngineer Watch:
4
6
38
Did you know? You can earn your doctorate while you work, as a research collaboration between your employer and Hopkins Engineering. Learn more at our next info session, Wed, Oct 22. Register:
applygrad.jhu.edu
0
0
1
JHU CS is ready for #ICCV2025! 🌺 Check out what our computer scientists are presenting at next week’s @ICCVConference in this thread 🧵: (1/10)
1
4
10
A team led by Professor Howard Katz and postdoctoral fellow Sasikumar Mayarambakam (pictured) has developed a wearable sensor that can detect acetone in breath—a possible early warning sign of diabetes—and sniff out harmful chemicals in the air. Read: https://t.co/6ymcRbUePt
0
1
1
A team of #HopkinsBME undergraduate students won the Krosnick Prize at the National Institute of Health’s 2025 DEBUT Challenge for a new device that helps doctors recognize hemorrhagic shock in pediatric patients. Congrats to team ShockSense! 👏 Read: https://t.co/ETZiKe7KZB
0
2
3
Using ML to identify neural firing patterns, @JHUBME Asst Prof Annie Kathuria can distinguish schizophrenia from bipolar disorder with 92% accuracy in lab-grown brain organoids—a breakthrough that could end the trial-and-error approach to medication.
engineering.jhu.edu
Using machine learning, Johns Hopkins researchers identified healthy and unhealthy patterns based on electrical activity
0
1
1
In @NatureComms: @JHUCaSE Asst Prof Hao (Frank) Yang's SafeTraffic Copilot, an LLM adapted to predict car crash outcomes as a reasoning task. Outperforms prior models by 33–46%, reveals top risk factors and supports smarter, safer interventions.
engineering.jhu.edu
SafeTraffic Copilot, created by Johns Hopkins team, aims to provide experts with crash analyses and predictions to reduce the rising number of fatalities and injuries that happen on U.S. roads each...
0
0
0
Nearly all results on multi-agent systems rely on 𝗼𝗳𝗳-𝘁𝗵𝗲-𝘀𝗵𝗲𝗹𝗳 models stitched together only at inference time. @jackjingyuzhang’s work breaks that mold: 𝗷𝗼𝗶𝗻𝘁𝗹𝘆 𝘁𝗿𝗮𝗶𝗻𝗲𝗱 multi-agent systems, specifically for improving safety.
💃New Multi-Agent RL Method: WaltzRL💃 📝: https://t.co/KE8dM9kX1r - Makes LLM safety a positive-sum game between a conversation & feedback agent - At inference feedback is adaptive, used when needed -> Improves safety & reduces overrefusals without degrading capabilities! 🧵1/5
0
3
15
Calling all JHU graduate students! IAA is hosting a lightning talks event on Tuesday, October 28 from 3:30-6 p.m. that will feature brief talks on topics related to AI and assurance. Learn more and submit a talk topic:
iaa.jhu.edu
No data
Save the date! IAA is hosting a Graduate Student Lightning Talks event on Tuesday, October 28 from 3:30-6 p.m., which will highlight short talks on topics related to #AI and assurance. For additional details and to submit a talk topic visit: https://t.co/5EUHfsFOPt
0
1
1
Johns Hopkins has been ranked No. 16 globally in the 2026 @timeshighered World University Rankings based on 18 key performance indicators in five categories.
hub.jhu.edu
The publication evaluates nearly 2,200 colleges around the world based on 18 key performance indicators in five categories: teaching, research environment, research quality, industry, and internati...
0
4
20
BDP Jim Bellingham discusses his new book, How Are Marine Robots Shaping Our Future?, on @shelleyrirwin's @WGVUPublicMedia show. Listen to the full conversation here: https://t.co/WkKjuV55cq
wgvunews.org
James Bellingham. PhD, shares his work.
0
2
1
Happening now at Impact Maryland, some of the Baltimore region's most visionary leaders and engaged community thinkers are gathering for a day of collaboration and forward-thinking dialogue. #ImpactMaryland
1
2
3
We introduce WaltzRL🎶, a multi-agent RL framework that treats LLM safety as a positive-sum game between conversation & feedback agents. It strikes an elegant balance between helpfulness & harmlessness, boosting safety & reduces overrefusals without degrading capabilities!
💃New Multi-Agent RL Method: WaltzRL💃 📝: https://t.co/KE8dM9kX1r - Makes LLM safety a positive-sum game between a conversation & feedback agent - At inference feedback is adaptive, used when needed -> Improves safety & reduces overrefusals without degrading capabilities! 🧵1/5
1
17
75
What is RNA velocity in situ? ChatGPT kept getting it wrong... So I made this video explaining how RNA velocity in situ infers gene expression dynamics by distinguishing nuclear vs. cytoplasmic expression in spatial transcriptomics data: https://t.co/IyvbNrZNBM
#AcademicTwitter
3
44
293
Mark Dredze (@mdredze) has been named inaugural director of our Data Science and AI Institute. A member of the @JHUCompSci faculty since '09, he is a pioneer in the application of AI for language analysis to public health and medicine.
engineering.jhu.edu
Dredze, a member of JHU's faculty since 2009, selected to lead institute dedicated to harnessing the power of artificial intelligence to translate data-driven discovery into real-world impact
0
10
58
We're inviting applications for a tenure-track faculty position (Assistant Professor, Associate Professor, or Full Professor) in the area of data science and artificial intelligence for materials design and innovation. Apply: https://t.co/FyvkcfHtBo
0
5
3
COLM 2025 was once of the best conferences I can recall ever attending. I heard this also from many in coffee chats and dinners. Thank you to all that put it together. I look forward to year three.
3
9
194
Exactly - just as @_beenkim et al, this work argues for interpretability agents that can model the understanding of listeners that interact with ML predictors. In that sense, interpretation of ML models should be targeted to specific world models to maximize utility (à la
@_beenkim @johnhewtt @NeelNanda5 We agree! Explanations should adapt to different users. Pragmatic reasoning formalizes this, as we show with Zhenzhen Wang, @PaulYiMD, @tianminshu, and @Jere_je_je. Our framework provides: ✅ Mutual mental model 🔁 Partial interaction 📜: https://t.co/y9V06q1IpK
0
1
1
.@JHUBME master's students won the top honor at the @ASMEishow for their invention, DINA—the Dynamic Infant Neurorehab Aid, a soft robotic suit created to help rehabilitate newborns with brachial plexus palsy (NBPP). Read more: https://t.co/HtrH93GZba
0
0
2