Learning for Medical Concepts Speaker: Shih-Han Lo Advisor: - - PowerPoint PPT Presentation

▶
learning for medical concepts
SMART_READER_LITE
LIVE PREVIEW

Learning for Medical Concepts Speaker: Shih-Han Lo Advisor: - - PowerPoint PPT Presentation

Jan 31, 2023 298 likes •544 views

Multi-layer Representation Learning for Medical Concepts Speaker: Shih-Han Lo Advisor: Professor Jia-Ling Koh Author: Edward Choi, Mohammad Taha Bahadori, Elizabeth Searles, Catherine Coffey, Michael Thompson, James Bost, Javier Tajedor-Sojo,

slide-1
SLIDE 1

Speaker: Shih-Han Lo Advisor: Professor Jia-Ling Koh Author: Edward Choi, Mohammad Taha Bahadori, Elizabeth Searles, Catherine Coffey, Michael Thompson, James Bost, Javier Tajedor-Sojo, Jimeng Sun Date: 2017/10/31 Source: KDD ’16

1

Multi-layer Representation Learning for Medical Concepts

slide-2
SLIDE 2

Outline

2

 Introduction  Method  Experiment  Conclusion

slide-3
SLIDE 3

Introduction

3

 Motivation

Ars Technica

slide-4
SLIDE 4

Introduction

4

 Purpose

 Learn interpretable representations.  Enable clinical applications to offer more than just

improved performances.

slide-5
SLIDE 5

Introduction

5

 Framework Electronic Health Records (EHR)

Code- and visit-level representations

Med2Vec

Input Output Proposed algorithm

  • r method
slide-6
SLIDE 6

Introduction

6

slide-7
SLIDE 7

Introduction

7

 EHR structure

 The set of all medical codes:  Sequence of visits: where  The goal of Med2Vec is to learn two types of

representations:

 Code representations  Visit representations

slide-8
SLIDE 8

Outline

8

 Introduction  Method  Experiment  Conclusion

slide-9
SLIDE 9

Method

9

 Med2Vec architecture

slide-10
SLIDE 10

Method

10

 Learning from the visit-level representation

 We minimize the cross entropy error as follows:

slide-11
SLIDE 11

Method

11

 Learning from the code-level representation

 The code-level representation can be learned by

maximizing the following likelihood.

−

slide-12
SLIDE 12

Method

12

 Unified training

Function (3) Function (2)

slide-13
SLIDE 13

Method

13

 Interpretation of learned representations

 Code representations

 Non-negative matrix factorization (NMF) 

 Visit representations

slide-14
SLIDE 14

Method

14

slide-15
SLIDE 15

Outline

15

 Introduction  Method  Experiment  Conclusion

slide-16
SLIDE 16

Experiment

16

 Datasets

slide-17
SLIDE 17

Experiment

17

 Evaluation strategies

 Code representations

 Qualitative evaluation by medical experts  Quantitative evaluation with baselines: NMI

 Visit representation

 Predicting future medical codes  Predicting CRG level  Baselines: One-hot+, SA, Skip-gram+, GloVe+

slide-18
SLIDE 18

Experiment

18

 Results

slide-19
SLIDE 19

Experiment

19

 Results

slide-20
SLIDE 20

Experiment

20

 Results

slide-21
SLIDE 21

Experiment

21

 Results

slide-22
SLIDE 22

Outline

22

 Introduction  Method  Experiment  Conclusion

slide-23
SLIDE 23

Conclusion

23

 We proposed Med2Vec for learning lower

dimensional representations for medical concepts.

 Med2Vec incorporates both code co-occurrence

information and visit sequence information of the EHR data.