Sketchy’s Post

We are thrilled to announce that JMIR AI: The Leading Journal for Real-World AI Applications in Medicine has been selected for inclusion in the Web of Science™. Passing the rigorous Editorial Evaluation ensures that JMIR AI is now eligible to receive its inaugural Journal Impact Factor (JIF) as early as 2025. This recognition underscores the journal's commitment to advancing applications of AI in health settings and reflects the incredible dedication of our editorial board, authors, and reviewers. Congratulations to Editors-in-Chief Khaled El Emam, PhD, and Brad Malin, PhD, and to all contributors who made this milestone possible! Explore some of our recent articles: Using Conversational AI to Facilitate Mental Health Assessments and Improve Clinical Efficiency Within Psychotherapy Services: Real-World Observational Study: https://hubs.la/Q02YYL1X0 Adolescents’ Well-being While Using a Mobile Artificial Intelligence–Powered Acceptance Commitment Therapy Tool: Evidence From a Longitudinal Study: https://hubs.la/Q02YYJB10 Developing Ethics and Equity Principles, Terms, and Engagement Tools to Advance Health Equity and Researcher Diversity in AI and Machine Learning: Modified Delphi Approach: https://hubs.la/Q02YYJSv0 Forecasting Artificial Intelligence Trends in Health Care: Systematic International Patent Analysis: https://hubs.la/Q02YYMny0 Assessing Elevated Blood Glucose Levels Through Blood Glucose Evaluation and Monitoring Using Machine Learning and Wearable Photoplethysmography Sensors: Algorithm Development and Validation: https://hubs.la/Q02YYJ8d0 🔗 View the full announcement: https://hubs.la/Q02YYJ3x0 #WebOfScience #JournalImpactFactor #AIinHealth #DigitalHealth #JMIRAI #ScholarlyPublishing

I watched a panel event - very accessible to me as a non medical person- and the issues raised are interesting per se but the parallels and wider implications for anyone thinking of the future of education and the role of AI were also there: *Data availability and quality, particularly with electronic health records. *The need for new business models to support preventative AI applications. *The importance of addressing human behaviour and encouraging the adoption of AI tools. It was interesting to see how these challenges mirror broader trends in AI adoption across various sectors. Successful integration of AI into healthcare requires us to think beyond just technology and consider the human and systemic factors at play. The same is clearly true in all the other disciplinary areas at our universities. https://lnkd.in/eNR7PJB9

STFM's Artificial Intelligence and Machine Learning for Primary Care Curriculum (AiM-PC) aims to equip learners with the skills needed to be engaged stakeholders, use #artificialintelligence #AI and #machinelearning in their practice, and ensure responsible and ethical use of AI/ML. Developed for medical students, primary care residents, and practicing primary care clinicians, the curriculum will be released in four parts from now through spring 2025, beginning with the Interview With an Innovator video series. Part 1 of the interview series, available now, features: Steven Lin, MD, Stanford University; Robert Wachter, MD, University of California, San Francisco; and Katie Link, product manager at NVIDIA and medical student at Icahn School of Medicine at Mount Sinai. Watch the interviews on STFM's website: https://lnkd.in/gTa6stJ5

After just one term of using Nectir AI Course Assistants, students and faculty at Los Angeles Pacific University reported a 20% increase in GPAs, a 13% rise in average final scores, a 36% boost in intrinsic motivation to learn, and a 13% improvement in self-efficacy. 🌎 This is what the campus of the future looks like in practice. Nectir AI creates a 24/7 support ecosystem for students and faculty in order to close critical equity gaps and create an equal playing field for everyone to succeed. ✅ See the full LAPU case study here and book a demo with our team to join the thousands of students, faculty, and staff using Nectir AI today: https://lnkd.in/gQSZcPVp

🚀 Excited to share that the ARMA - NL Working Group on Generative AI is gaining momentum! Last week, we hosted the first in-person meeting at UMC Utrecht. I'm working alongside 24 colleagues from universities, university medical centers, and universities of applied sciences across the Netherlands. Our belief is simple: GenAI will transform how we work in research management and administration (RMA). Together, we're on a mission to empower RMAs to not only adapt but actively shape the future of our profession. What's next? For 2025, we're already planning: 📚 Training sessions and workshops 🤝 Knowledge exchange events 🔍 Mapping of AI initiatives Stay tuned for more updates! #ResearchManagement #AI #GenerativeAI #Innovation #HigherEducation #ARMANL #RMA

Crocus is excited to announce its collaboration with the California School Boards Association on its most recent endeavor: the creation of an Artificial Intelligence Taskforce. The goal of this initiative is to explore the implications of the widespread adoption of AI in education from the standpoint of policy and best practices, impacting teaching, learning and operational efficiency in California public schools. Through the AI Taskforce we open dialogue with key stakeholders from 28 California school districts and county offices including board members, superintendents, administrators, educators and PreK-12 IT experts. AI has revolutionized many sectors during the last year. Therefore, we need to get involved to collectively define the best way to leverage AI to innovate and generate positive impact and minimize harm in the public education sector. #EducationInnovation #AIInitiative #AIforEducation #TechForImpact #AITaskforce

I'm thrilled to announce that my first-author paper, "Reinforcement Learning in Dynamic Treatment Regimes Needs Critical Reexamination," has been accepted as a spotlight paper at ICML 2024! 🎉🎉🎉 This achievement is particularly significant given the competitive landscape of ICML, where only 3.5% of submissions were selected as spotlights. Out of 9,473 total submissions, our paper is among the 191 spotlight-designated works! In our paper, we delve into the challenges and opportunities of applying offline reinforcement learning (RL) in dynamic treatment regimes (DTRs) within healthcare. Through an extensive case study involving over 17,000 evaluation experiments on a Sepsis dataset, we uncover critical insights about the variability and performance of RL algorithms in different contexts. Our findings underscore the need for more robust policy evaluation and algorithm development in future DTR research. I want to extend my heartfelt thanks to Prof Tingting Zhu and Dr Yangchen Pan, whose support and guidance were instrumental in this accomplishment. You can read the paper here: [Position: Reinforcement Learning in Dynamic Treatment Regimes Needs Critical Reexamination](https://lnkd.in/eq7ZzTUJ) and access our code here: [GitHub - ReassessDTR](https://lnkd.in/ew2KCfpb). Looking forward to further discussions and advancements in this critical area! See you in Vienna! #ICML2024 #ReinforcementLearning #DynamicTreatmentRegimes #HealthcareAI #Research #AI

🥳 Proud to share with you my latest research article as a preprint now online on #bioRxiv! "Dopamine reveals adaptive learning of action representation" https://lnkd.in/eW6y2YHQ 🎯 In this research article, I delve into the theories of decision making, reinforcement learning and cognitive representations, through experiments on mice behavior and dopamine measures. Like Pavlov’s dog associating a bell 🔔 with food 🍖 , humans and other animals continuously try to infer associations between predictive events and outcomes (good ✅ or bad ❌ ). Through experience, our brain 🧠 learns to assign predictive values to environmental items, building subjective preferences and updating our mental models of the world 🤔. This process, called reinforcement learning, involves computing a "Reward Prediction Error" (RPE, i.e. a comparison between what was expected and what was really obtained), signaled by the neurotransmitter dopamine, to adjust behavior. Leveraging a decision game where mice 🐭 have to solve different reward delivery rules to gain rewards 🏁, I discuss in this paper how different items or features (e.g., a bell ring or a red light) can hold different predictive values depending on context, and how analyzing dopamine can reveal which features our brain uses to build expectations and predict outcomes. Moreover, I show that when changing context (e.g., from rule A to rule B), dopamine signal also reveals a change in the internal model of the world (e.g. now, it is the red light that matters to predict a reward and guide decision). This work has been conducted in #ESPCI #BrainPlasticityLab #PSL and #SorbonneUniversités, with funding from the #FondationRechercheMédicale. On top of years of experiments and several presentations 👨🏫 in world-class congresses (#Dopamine2022, #NeuroFrance2022, #FENS2024), it also involves a great deal of data analysis 📈, statistics and AI to extract such exciting results! I thank my supervisor #PhilippeFaure and all my colleagues who were of great support to bring this project that far! 🚀 If you want to go deeper, please feel free to dive into the texts and figures, or message me for more details.

We have been biting our lips but today can announce the Ryght Research Network, a global network of clinical trial sites, with USC as our first major academic medical center to join the network. There will be more announcements soon, but our vision from day one has been that if all participants in a connected ecosystem are using the same #generativeAI, there can be a real time exchange of information and value. Soon, the days of sending out emails (or, god forbid, faxes) to disconnected sites to find out if they have the patients and capabilities for a clinical trial will be long gone. You’ll just know :-). These and many other use cases for sites, sponsors and CROs enable a real time, AI-driven system that’s set to transform the industry. Welcome Keck School of Medicine of the University of Southern California to the Ryght Research Network.

I prefer to deal with LLM hallucinations over real-world illusions... AI hallucinates, yes but without intent—reflecting the data it’s trained on. In contrast, real-world illusions—biases, inequities, and constructed narratives—sustain systemic issues. In HUMBLE we aim to bridge these worlds: https://lnkd.in/dJ-YmGFb Can AI expose human bias while humans mitigate AI bias? Can it help us see through the illusions we create? #BigData RREAL (University College London) #UCL UCL Institute of Health Informatics  Wellcome Trust Julia Ive