1 / 36

Lawrence Carin Duke University

The Coming AI Disruption of Healthcare. Lawrence Carin Duke University. So, What is Deep Learning and AI?. Learning a Predictive Model Based on Labeled Data. y , associated label 0/1. x , data for a subject. Training Set. Linear Predictive Model. Logistic Regression. ].

bbarnard
Télécharger la présentation

Lawrence Carin Duke University

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. The Coming AI Disruption of Healthcare Lawrence Carin Duke University

  2. So, What is Deep Learning and AI?

  3. Learning a Predictive Model Based on Labeled Data y, associated label 0/1 x, data for a subject

  4. Training Set

  5. Linear Predictive Model

  6. Logistic Regression ] Large implies that it is likely

  7. Learning Model Parameters ]

  8. Generalization of Logistic Regression: Learned Features Probability of K latent processes/features

  9. Extended Logistic Regression Probability of particular outcome Probability of K latent processes May be viewed as logistic regression on K latent features, rather than directly on the N components of raw data

  10. Going Deeper Probability of J Layer-2 latent processes Probability of K Layer-1 latent processes

  11. Probability of particular outcome Probability of J Layer-2 latent processes Probability of K Layer-1 latent processes

  12. Increasing Degree of Specificity and Feature Sophistication

  13. Multilayered Perceptron: Neural Network

  14. The Trade: Bias vs. Variance ?

  15. The Seasons of Neural Networks Multi-Layered Perceptron 1960

  16. The Seasons of Neural Networks Multi-Layered Perceptron BackPropagation 1960 1986

  17. The Seasons of Neural Networks Multi-Layered Perceptron BackPropagation Convolutional Neural Network 1960 1986 1989

  18. The Seasons of Neural Networks Multi-Layered Perceptron BackPropagation Convolutional Neural Network Neural Nets in the Wild 1960 1986 1989 1990–1994

  19. The Seasons of Neural Networks Long Short-TermMemory Multi-Layered Perceptron BackPropagation Convolutional Neural Network Neural Nets in the Wild 1960 1986 1989 1990–1994 1995

  20. The Seasons of Neural Networks More Neural Nets In the Wild 1998–2005

  21. The Seasons of Neural Networks More Neural Nets In the Wild Banishment 1998–2005 2005–2010

  22. The Seasons of Neural Networks More Neural Nets In the Wild Rename:Deep Learning Banishment 1998–2005 2005–2010 2010–

  23. The Seasons of Neural Networks More Neural Nets In the Wild Rename:Deep Learning CNN+GPU+ Image Net Banishment 1998–2005 2005–2010 2010– 2013

  24. The Seasons of Neural Networks More Neural Nets In the Wild Rename:Deep Learning CNN+GPU+ Image Net Alpha Go Banishment 1998–2005 2005–2010 2010– 2013 2015

  25. Those Who Cannot Learn from History Are Doomed to Repeat It ?

  26. Deep Learning + Transfer Learning Convolutional Layers Convolutional Layers Fully Conn. Layers Fully Conn. Layers ImageNet (public image database with 1.2M+ images) Categories Weights Confocal Images Scores

  27. AI in the Clinic: Medical-Image Analysis Ophthalmology Pathology Deep Learning

  28. AI for the Basic Sciences of Health: Neuroscience K. Dzirasa, L. Carin, et al., “Brain-wide Electrical Spatiotemporal Dynamics Encode Depression Vulnerability,” Cell, 2018

  29. Value-Based Care

  30. Enhancing Patient-Doctor Experience with AI • Clinician burn-out and depression significant national issue • The EHR has imposed significant data entry on doctors • Not why people become doctors, also serious degradation • of patient experience • Opportunity for AI: • Audio processing to transcribe all MD-patient verbal • communications • Use natural language processing to automate data entry • Machine learning to assist MD in diagnosis & treatment

  31. AI to Automate Answering Routine Questions of Clinicians

  32. The Coming AI Disruption of Healthcare • AI has had many seasons, this time spring/summer seems here to stay • Many demonstrations of utility, in the real world • Significant disruption in motion for healthcare, that will be accelerating: • - Analysis of all forms of clinical images: radiology, pathology, ophthalmology, etc. • - Health-system optimization • - Patient-clinician interactions • - Basic sciences for health

More Related