Omar Sanseviero (@osanseviero)

MedGemma Impact Challenge가 종료됐다. 850개가 넘는 팀이 참여해 서아프리카 국가들의 다국어 비정형 임상 관찰 기록을 활용하거나 피부 변화 추적 등 의료 혁신 프로젝트를 제작했다. 헬스케어 AI 응용 사례와 모델 활용 가능성을 보여준 의미 있는 대회였다.

https://x.com/osanseviero/status/2037270697160958206

#medgemma #healthcareai #aiapplication #multimodal #clinicalai

Omar Sanseviero (@osanseviero) on X

The MedGemma Impact Challenge has concluded 👀 Over 850 teams built great projects aiming to transform health, from enabling west african states to transform multilingual unstructured clinical observations to skin changes tracking. Lots of amazing projects

X (formerly Twitter)

Why are doctors warming up to generative AI? Pitt researchers explore how large language models could boost patient care, streamline workflows, and raise ethical questions. Discover the potential benefits and challenges of AI tools in clinical practice. #GenerativeAI #LargeLanguageModels #ClinicalAI #MedicalEthics

🔗 https://aidailypost.com/news/pitt-examines-why-doctors-might-embrace-generative-ai-clinical

Oracle Health Clinical AI Agent now drafts labs, meds, and follow-up orders while it listens to your visit. How this shift cuts clicks and burnout risk for doctors. Read the full breakdown #AdwaitX #ClinicalAI #OracleHealth #news
https://www.adwaitx.com/oracle-health-clinical-ai-agent-order-creation/
Oracle Health Clinical AI Agent Learns to Write Orders While You Talk to Patients

Discover how Oracle Health Clinical AI Agent drafts clinical orders and notes, helping clinicians cut documentation time and focus on patient care. AdwaitX

AdwaitX News

Clinical AI can ace ROC AUC and still create harm: missed cases, alert fatigue, and workflow friction.

Learn evaluation that matches bedside decisions: Calibration, cost‑sensitive utility, and Decision Curve Analysis (net benefit).

Read → https://codelabsacademy.com/en/blog/designing-clinical-ai-evaluation-protocols-utility-harm?source=mastodon

#ClinicalAI #MachineLearning #HealthTech

Designing Clinical AI Evaluation Protocols Beyond ROC AUC

Learn to evaluate clinical AI beyond ROC AUC with calibration, cost-sensitive metrics, and decision curves to measure utility, harm, and workflow impact.

A new MIT study reveals that large language models can unintentionally memorize details from de‑identified patient notes, exposing a privacy gap in clinical AI. Researchers led by Arash Tonekaboni show why stricter safeguards are needed. Read on to see the implications. #MITAI #ClinicalAI #DeIdentifiedData #MemorizationRisk

🔗 https://aidailypost.com/news/mit-study-probes-memorization-risk-clinical-ai-deidentified-data

Adaptive sampling #WGS is moving genomics from lab bench to bedside. In our latest work, we deliver variant, structural & methylation insights on @nanopore's PromethION platform with molecular signals emerging in minutes, not days. Watch this space 👀
#clinicalai #genomicsforhealth #nanopore #multiomics

https://www.medrxiv.org/content/10.1101/2025.10.02.25336569v1

🩺🤖 What hidden stories do electronic health records tell about aging?

🔗 Identifying health conditions in older adults in textual health records using deep learning-based natural language processing. Computational and Structural Biotechnology Journal, DOI: https://doi.org/10.1016/j.csbj.2025.09.009

📚 CSBJ Smart Hospital: https://www.csbj.org/smarthospital

#AIinHealthcare #DigitalHealth #NLP #MachineLearning #DeepLearning #EHR #HealthData #HealthyAging #MedTech #HealthTech #ClinicalAI #Geriatrics #PrecisionHealth