250 likes | 275 Views
Join our mission to advance AI in medicine at UCLA, transforming patient care through data science. Explore our research areas and partnerships to revolutionize healthcare.
E N D
Departmental Goals • Become the leading program in data science applied to biomedical data • Transform patient care in the UCLA Health System by leveraging advances in AI and Machine Learning
Our plan • Expand research scope and educational programs • Develop AI in Medicine @ UCLA Programs • Leverage UCLA’s leadership in Data Sciences:Engineering School Affiliation
Expanded Scope Eleazar EskinJason ErnstBogdan PasaniucWei Wang Eran HalperinJessica LiNoah ZaitlenSriram Sankararaman New departmental faculty in Computational Genomics: Jointly appointed with: StatisticsAnesthesiologyBiological ChemistryComputer Science Human GeneticsNeurologyPathology Growth Areas of Data Science in Biomedicine • Clinical Machine Learning • Computer Vision in Medicine Leverage existing strength in Mathematical Modeling
What is AI in Medicine? • Impact care: • Identify high risk cases. • Improve speed and accuracy of diagnosis. • Predict treatment outcomes and side effects. • Reduce costs. • Requires research in AI and research in Medicine
What is NOT AI in Medicine? • Analysis of small numbers of patients and/or small numbers of features • Utilizing “Machine Learning Software” • Hiring “Data Scientists” • Information Technology
In Partnership with UCLA Health • Electronic Health Records (2M+ records) • Genomic Data (25,000 DNA genotypes + 750 week) • Waveforms (30,000 patients per year) • Images (1,500 imaging studies per day) • more…
AI in Medicine @ UCLALed by EranHalperin • CompMed/Anesthesiology Collaboration: Preoperative predictions of in-hospital mortality using electronic medical record data • Effort led by Eran Halperin/Ira Hofer. • Used EHR data from ~50,000 patients going through surgery. • Applied machine learning techniques to predict preoperative risk for mortality (AUC = 0.92). • Saves the time of ASA Physical Status Classification determined by the anesthesiologsits (AUC = 0.86) • Next step: Improve care and reduce costs by saving anesthesiologists time. • Creating a network of collaborations CompMed+ Clinical Departments (OBGYN, Neurology, Anesthesiology)
AI in Medicine @ UCLA CROWDED AREA ENABLE CLINICAL SCIENCE WITH HEALTH DATA PARTNERSHIPWITHCLINICAL CHAIRS
Computational Medicine Programs • Blockchain and Secure Computation in Medicine: Leverage transformative technology breakthroughs in cryptography • Effort led by SriramSankararaman, Amit Sahai • Develop technology for analysis of genomic and medical data while preserving privacy and data sharing policy. • Builds upon foundational research @ UCLA: Secure Computation (Sahai), Differential Privacy in Genetics (Sankararaman), Private Genome Analysis (Sahai, Eskin) • Computational Genomics in the Health System:Computational methods for combining genomic and health system data • Effort led by Jessica Li, Jason Ernst, BogdanPasaniuc • Builds upon foundational research @ UCLA in Computational Genomics • Testbeds are UCLA Health System through AtLAs project and Depression Grand Challenge • Tech Transfer to Industry: Commercialization and Software Licensing
UCLA Computational Medicine is a Hub in the Scientific Community
Industry Partners • Industry Partners will translate discoveries beyond UCLA Health System • Current Partnerships • Amazon • We are looking for industry partners to: • Develop AI in Medicine Technology • Answer Medical Questions using UCLA Health Data (e.g., stratifying patients receiving drug to reduce side effects)
Academic Medical Center Partnerships • Goal: Establish Formal Collaborative Relationship with Health Systems which have Genomic Patient Data • Replication across Health Systems • Meta-Analysis (e.g. NIDA Opiod Consortium) • Share expertise and collaborate • Current Partnerships in Progress • Vanderbilt (Nancy Cox) • University of Michigan (Michael Bohenkhe) • Mt. Sinai (Eimear Kenney)
Scientific Community Leadership • Computational Genomics Summer Institute • 50 Faculty and 150 participants spent 1 week to 1 month at UCLA in 2018 • External Affiliate Member Program • Incorporate Scientific Community Expertise in Analyzing UCLA Health Data • Within UC Expertise • Outside UC Expertise
UCLA’s Potential • Best group of faculty in Computational Genomics + Machine Learning in the Country • Strong Computer Science / Engineering • International leaders in Machine Learning, Data Mining, Computer Vision and related areas. • Strong Medical School • Excellent and large Health System -> unique data. • Single Campus • A history of interdisciplinary activities.
UCLA Computational Medicine • New DGSOM Department (formerly Biomathematics) • Department Goals: • Become the leading program in data science applied to biomedical data • Transform patient care in the UCLA Health System by leveraging advances in AI and Machine Learning • Department Scope: • Computational Genomics • Machine Learning on Clinical Data • 8 Faculty joined Biomathematics in 2018 • Ongoing FTE Faculty Search • Provides bridge to Data Science Expertise on the Main Campus • Computer Vision on Medical Images • Mathematical Modeling