John Humphrey Plummer Professor of Machine Learning, AI, and Medicine at University of Cambridge
Ready for the next big leap in making AI truly accessible? We are excited to release #CliMB, a no-code AI-enabled partner for clinical predictive modelling! With CliMB, you can build predictive models using natural language. CliMB supports data exploration, engineering, model building, and interpretation—enabling clinician scientists to utilise cutting-edge tools in the fields of data-centric AI, AutoML, and interpretable ML. This proof of concept is a huge step towards breaking barriers and #empowering clinician scientists to build predictive models using cutting-edge tools! You can read our paper here: https://lnkd.in/eC2M5WMw You can download CliMB here: https://lnkd.in/eeWWU_uV We recently discussed CliMB extensively with clinical researchers during a #RevolutionizingHealthcare session. You can watch the full episode on YouTube to hear their insights and watch our demonstrations of the tool: https://lnkd.in/ekGefg3d
Dear Mihaela van der Schaar,Evgeny Saveliev, Tim Schubert, Thomas Pouplin and Vasilis Kosmoliaptsis MD, FRCSEng, PhD(Cantab): I read your paper and I love that you explore the field of making AI more accessible for domain experts in medicine. A thought: On p. 4 you explain that AutoML does "not deliver on its original purpose of democratizing ML", especially because it is no "off-the-shelf ML solutions for their problems" and because "most AutoML tools today require technical understanding and skills (e.g., the ability to use Python packages like AutoPrognosis [13])." While I agree that AutoML is not perfect and still needs to be more human-centered as argued by Marius Lindauer, Florian Karl et al. I still personally believe that this judgment is too negative. There is quite some research on AutoML's utility for non-experts. Me and Dr. Marc Schmitt specifically wrote this Q1 paper which shows how no code AutoML solution can be a solution: https://doi.org/10.1080/10447318.2024.2425454 I think that including this perspective in "related research" would create a more realistic image of the real potential of AutoML for domain experts. I hope that you interpret this comment as my effort to improve your research 👍
I was also trying to build something like this for AutoML. Amazing work!
Impressive
Looks great
Clinician, Academician & Advocate, Digital Health & AI Medicine | FIFA Certified Sports Physician | CMgr CCMI (CMI UK) | MBA | MBBS | PhD Scholar (Digital Health & AI Medicine) | Royal Watchmaker | CEO & Horologer, MHKL
1moDear Prof Mihaela van der Schaar, may I know the basic system requirements to run CliMB for a small-scale research unit? Thank you.