What is AI? Saeed Ahmed MD Disclosures This presentation reflects - - PowerPoint PPT Presentation

What is AI?

Saeed Ahmed MD

Disclosures

• This presentation reflects the views of the author and not of any
• rganization.
• Independent consultant currently on contractual assignment with

Biogen

• Other financial disclosures: none.

Art rtificial In Intelligence

Narrow vs. . General

What do these guys think about Artificial General Intelligence?

What is Machine Learning?

“Curve Fitting” [multivariate, large datasets, fast computation]

Judea Pearl (2019) SamHarris.org Chancellor’s Professor Emeritus, Computer Science, UCLA

A machine is said to learn from experience E with respect to some class

• f tasks t and performance measure P if its performance at tasks in t, as

measured by P, improves with experience E

Mitchell, Tom. (1997) Machine Learning

• E. Fredkin University Professor

School of Computer Science Carnegie Mellon University

P(t) E

Data

Structured and Unstructured

Learning

Unsupervis ised and Superv rvis ised

Learning

Unsupervis ised and Superv rvis ised

Learning

Unsupervis ised and Superv rvis ised

Recruitment based on referral and local databases Site-based Data-Collection Pre- programmed Rules of Analysis based

• n SAP

Tables and Graphs Low info transfer Smart Matching Social Media Site-less data collection New data types Multimodal data fusion Machine Learning Iterative improvements to model Hi info transfer

Clinical Research - Traditional vs. New Approaches

Clinical Research Models

Old and New

Handle with Care!

Data Output (‘prediction’)

Training/Deployment

Data

X-AI

Output (‘prediction’)

Black Box

Data

Output (‘prediction’)

Scalability

Practic ical Consideratio ions

Careful Sampling is Still Very ry Im Important

Set consisting of all possible data from all patients with schizophrenia T1 T2

• Thank you!