Note: I am writing at the event so please excuse any typos and these notes will contain my personal comments
Steve Murphy WebMD
1. Translational Science: Transforming Data to Knowledge to Improve Health. Rosemary Filart, MD MPH
What is Translational Science? Specifically t2, T3, T4
Advancing Research to Improve Human Health
What Data? Use BIG DATA? Many challenges
Data applications? Repeatable. Reproducible. Benefits clinicians, T&C
Making more informed decisions at the Point of Care. Research towards individualized, personalized medicine
Dissemination. Implementation and dissemination of info and research questions are vital.
2. Radical Reuse: Repurposing Yesterday’s Data for Tomorrow’s Discoveries. Lisa Federer MLIS, MA, AHIP, NIH
Radical reuse. Using shipping containers into luxury hotel rooms. Found objects into something new and useful
Facilitating Data reuse.
Taking analog expertise to the digital
Description: standardizing metadata
Discoverability: Data catalogs – Should be true for Data
Dissemination: Facilitating Sharing
Digital Infrastructure: Tools & Systems
Data Literacy: Training Scientists
Rethinking the Future of Data- Paradigm shift
3. Discovering Medical Knowledge with BTRIS Data. James Cimino MD, NIH Clinical Center
Clinical Data at NIH –
BTRIS in a nutshell. Stats are phenomenal
Limited Data Sets
Can select – Diagnosis, demographics, medications, lab tests
Opportunities for Data Mining
4. Discovering Data From BTRIS. Vojtech Huser MD, Ph.D.
1/3 Problem List
Meaningful use, Stage 1, Core Criteria
2/3 Problem List: Medication. Tons of medication data.
3/3 Problem List: Pharmacogenomics with whole exome
5. A TB or Not TB: Detecting TB in Remote Geographic locations. a whole bunch of people who I can’t list from NIH presented by Dr Stefan Jaeger
TB is a world health problem
Active in Kenya. Partnership with USAID and AMPATH
Portable X-ray machines on trucks and take to the remote villages then bring the patients to hospitals and clinics
Firefly Labeling Tool
6. New computational Intelligence. Jim DeLeo NIH
John von Neumann
What can I get the machine to do? What do you want the machine to do?
Computational Intelligence both artificial intelligence and machine learning.
Concerned with clustering and classification that sets the stage for hypothesis
Lots of computational tools.
Biomedical data mining
Heterogeneity. Heterogeneous – higher granularity
Ensembles and evolving ensembles
Clinical decision support system
Extreme Multidisciplinary Teaming
Current projects – running out of time so speeding through the topics
7. Deep Learning. Jonathan Simon
Machine Leaning. Subfield of AI.
Neural Networks algorithms
Deep Neural Networks. Characterized by the fact that they are Very Very Large, i.e. FB.
Biomedical Deep Learning is starting also.
Possible Future Uses.
Brain hurts. These people are FAR smarter than I will Ever be. Nice to know that they are using their talents for Good!!!!