Lequan Yu
@ylqzd2011
Followers
659
Following
404
Media
0
Statuses
56
Assistant Professor at @HKUniversity, Medical AI, Medical Image Analysis, AI for Healthcare / former postdoc at @StanfordMed @StanfordRadOnc
Hong Kong
Joined December 2015
As AI takes on agent roles in critical fields, reasoning failures raise risks. New studies from @james_y_zou of @StanfordMed and @ylqzd2011 of @HKUniversity show how reasoning goes off the rails.
spectrum.ieee.org
As AI takes on agent roles, flawed reasoning raises risks
0
6
13
📢 Deadline extended! Submit your paper to #ISBI2026 by 14 Nov 2025. 🔗 https://t.co/42t6EdTvTU
1
5
21
🚨 Last week to submit your paper to #ISBI2026! 🚨 IEEE ISBI 2026 is calling for original work in biomedical imaging: 🧠 AI & image analysis 🔬 Multimodal/multiscale imaging 💡 Computer-aided diagnosis & more! 📅 Deadline: Oct 31, 2025 📚 4-page papers published in IEEE Xplore
1
1
1
This is not a great news for Medical AI. ☹️ Shows that medical AI Agents built from multiple LLMs often look correct but think wrong. Even when they give the right diagnosis, most of them got there through broken teamwork. In over 68% of “successful” cases, the agents already
54
141
494
🚀 Excited for ISBI 2026! A huge THANK YOU to our new & returning supporters 💙 Join us as a sponsor & enjoy amazing benefits 🎁✨ Exclusive Diamond & Diamond Plus packages available 💎 📧 Contact: ckravec@conferencecatalysts.com #ISBI2026 #Sponsorship #Networking #Innovation
1
3
5
🚨 Call for Papers 🚨 We’re launching a collection in @npjDigitalMed on “Evaluating the Real-World Clinical Performance of AI”. Submit your work by June 2026 and help shape the future of clinical AI! #AI #DigitalHealth #MedicalAI
https://t.co/6PG5jG4b9h
nature.com
This Collection invites research on exploring how AI performs in real-world clinical settings, focusing on utility, safety, equity, and impact on healthcare.
0
3
3
⏰ Deadline approaching! Less than 1 month left to submit proposals for demos, tutorials, challenges, and special sessions at ISBI 2026! 📅 Submit by September 15, 2025. Don't miss it! 🔗 https://t.co/RpxGHk2JM2
#ISBI2026 @IEEEsps
0
2
4
We accept both 8-page-long papers and 4-page-long abstracts and welcome for the submission.
🚨 Call for submission at our workshop on Computer Vision for Automated Medical Diagnosis @ICCVConference in 🌴 Honolulu, Hawaii! Join us at #ICCV2025 to present your work on 🤖multimodal LLMs, 🧠 agents, & ⚖️ fair, reliable AI for healthcare and medicine. 🗓️ Deadline: June 21,
0
1
8
📣 Stay Tuned! HKU Summer Research Programme 2025 will soon open for application! Join us to establish a global research network and get a taste of doing your own research project from scratch! Application period: 2 Jan (10am HKT) - 3 Feb 2025 (5pm HKT) https://t.co/IUmNeLWcsm
0
4
4
Excited to share our work and views on AI-driven precision oncology.
AI vs cancer. Professor Lequan Yu is the director of Medical AI Lab at HKU. In this interview, part of our Conversations about AI series, Prof Yu explains how, in the near future, AI can revolutionize the treatment of cancer: by integrating patient’s medical information - scans,
1
2
19
Are you interested in studying PhD and MPhil programmes at HKU? Join us at the PhD & MPhil Info Day 2024 on October 17, 2024 to learn more! Register now: https://t.co/IHxq3Q3iLq Further details: https://t.co/aJzEqBS8HA
0
3
8
We propose a new framework for AI-enhanced multimodal imaging. Check our paper for more details.
Multi-sensor learning (MSL) enables information transfer across different imaging modalities—a novel framework and significant milestone in AI multimodal imaging! I hope we will see the reported MSL images in routine practice soon! Kudos to co-authors! https://t.co/vReTHPeO5I
0
1
21
🎉 Exciting news! We're gearing up for the #MICCAI2024 tutorial on 'Graph Learning in Medical Image Analysis (GraphMIA)'! 🚀 Join us as we dive deep into the cutting-edge applications of graph learning in the medical imaging field. #GraphMIA2024 🔗: https://t.co/WwniTbRYP1
0
2
10
A bit slow on X but still super excited to share our Hist2Cell, predicting fine-grained cell types (e.g., 80 subtypes in lung) solely from histology images. For less funded labs to go for spatial RNAs, you can observe a lot more just from H&E stains (1/4)
biorxiv.org
Histology images, with low cost, are unleashing great power of predicting cellular phenotypes in tissue, thanks to the emerging spatial transcriptomics serving as annotations. Recent efforts aimed to...
2
19
53
Thrilled to see our SpatialDM online w/ @ZhuoxuanHKU, Tianjie & Pentao. Try your large-scale spatial transcriptomic data now - it's scalable! The reviewers are wonderful - both reviewers always responded in ~2 weeks w/ constructive comments!
SpatialDM for rapid identification of spatially co-expressed ligand–receptor and revealing cell–cell communication patterns @YuanhuaHuang @NatureComms #BiotechNatureComms #spatialtranscriptomics #singlecell
0
3
21
The development of machine-learning systems for safer, robust and fairer outcomes should leverage fine-tuning, generalization, explainability and metrics of uncertainty. https://t.co/arToxY2HYJ
0
4
11
Co-organizers: @yang_yuzhe @giholste @kyrie0109 @pengyifan @ylqzd2011 @HaoChen_HKUST @p_tingying @hsharonx Atlas Wang @yuyinzhou_cs
@ICCVConference #ICCV2023 #ICCV #CV #Health #MedicalAI #ML #AI
0
1
1
📢 Call for Participants 📢 Join *CXR-LT*, a competition for multi-label, long-tailed diagnosis on chest X-rays! Top teams may present their solutions for publication at our #ICCV2023 CVAMD workshop ( https://t.co/UqY3yKsENi)! Join here: https://t.co/NStRTBbOzP
1
8
23
This week, @kyrie0109 from HKU will be joining us to talk about multi-granularity cross-modal alignment for generalized medical visual representation learning. Catch it at 10-11am PT this Thursday on Zoom! Subscribe to https://t.co/6oTrfQjUkr
#ML #AI #medicine #healthcare
0
2
8
HKU Summer Research Programme 2023 – Apply Now! 10-week intense research training programme with one-to-one mentorship for elite students around the world. Outstanding students may receive a conditional offer of HKU Presidential PhD Scholarship. Details : https://t.co/E6nrvttYbw
1
7
19