NLP approach for medical translation task
2 likes631 views
This document discusses how artificial intelligence and natural language processing can help address challenges in healthcare data management and standards. Currently, there are many healthcare data standards but they are complex and difficult for doctors to use. Artificial intelligence approaches like word embeddings can help by analyzing large amounts of medical text to learn relationships between similar clinical terms and concepts without explicit programming of rules. This has the potential to help with tasks like mapping between different clinical coding standards and enabling interoperability between different healthcare data sources.
![束One or two patient died per week in a
certain smallish town because of the lack
of information 鍖ow between the
hospitals emergency room and the
nearby mental health clinic損
[束Doing Data Science損, ONeil ]](https://image.slidesharecdn.com/healthcaredatascienceif-151221142021/85/NLP-approach-for-medical-translation-task-9-320.jpg)
![[Image Source]](https://image.slidesharecdn.com/healthcaredatascienceif-151221142021/85/NLP-approach-for-medical-translation-task-13-320.jpg)
![束It was the year when Udacity, Coursera and edX, the three
leading MOOC companies, took the education world by storm
and promised a lot損 [Huf鍖ngton Post]
束Many places offer MOOCs, and many more will. But
Coursera, Udacity and edX are the leading
providers.損 [NYTimes]](https://image.slidesharecdn.com/healthcaredatascienceif-151221142021/85/NLP-approach-for-medical-translation-task-22-320.jpg)
![All
you
need
is
love
[0.2, 0.11, 087, 0.9, , 0.2]
[0.1, 0,98, 01, 0.26, , 0.82]
[0.7, 0.22, 0.3, 0.1, , 0.45]
[0.5, 0.21, 0,67, 0.82,, 0.49]
[0.6, 034, 0.21, 0.45,, 0.2]](https://image.slidesharecdn.com/healthcaredatascienceif-151221142021/85/NLP-approach-for-medical-translation-task-26-320.jpg)
NLP approach for medical translation task
- 1. When Healthcare Meets Data Science Anastasiia Kornilova
- 4. The Medicine of the Future
- 9. 束One or two patient died per week in a certain smallish town because of the lack of information 鍖ow between the hospitals emergency room and the nearby mental health clinic損 [束Doing Data Science損, ONeil ]
- 10. 60% of US doctors still use paper medical records
- 11. Lets create our own EHR standard
- 12. Patient gender Code Male 0 Female 1 Patient gender Code Male 1 Female 0 Patient gender Code Male M Female F Unknown U Lets code gender Standart A Standart B Standart C x x
- 13. [Image Source]
- 14. There 5 important data standards ICD - diagnostic, billing, world-wide CPT - procedures, billing, US-speci鍖c, classi鍖cation LOINC - lab tests and observations, world-wide NDC - medication, US-speci鍖c, classi鍖cation SNOMED - medicine
- 15. and a lot of custom dictionaries
- 16. Even within one data standard: ICD-9 174 malignant neoplasm of female breast 174.1 malignant neoplasm of central portion of female breast ICD-10 C50 malignant neoplasm of breast C50.1 malignant neoplasm of central portion of breast C50.111 malignant neoplasm of central portion of right female breast C50.112 malignant neoplasm of central portion of left female breast
- 17. You have to be a doctor to handle them
- 18. Problem summary Source 1 Source 2 Source N medicine expertise a lot of (expensive) hours Knowledge
- 20. Arti鍖cial Intelligence Way Feed a lot of medical texts to 束medical doctor損 Use NLP power Make it unsupervised
- 21. Key idea: 束Semantically similar words occurs in similar contents損 Harris, 1954 束You shall know a word by the company it keeps損, Firth, 1957
- 22. 束It was the year when Udacity, Coursera and edX, the three leading MOOC companies, took the education world by storm and promised a lot損 [Huf鍖ngton Post] 束Many places offer MOOCs, and many more will. But Coursera, Udacity and edX are the leading providers.損 [NYTimes]
- 23. Distributed Vectors Representation Two layer neural network Input: text corpus Output: set of vectors Group the vectors of similar words together in vector space (detects similarities matematically)
- 24. Predict a word using content All you need love is
- 26. All you need is love [0.2, 0.11, 087, 0.9, , 0.2] [0.1, 0,98, 01, 0.26, , 0.82] [0.7, 0.22, 0.3, 0.1, , 0.45] [0.5, 0.21, 0,67, 0.82,, 0.49] [0.6, 034, 0.21, 0.45,, 0.2]
- 33. ICD-9 174 malignant neoplasm of female breast 174.1 malignant neoplasm of central portion of female breast ICD-10 C50 malignant neoplasm of breast C50.1 malignant neoplasm of central portion of breast C50.111 malignant neoplasm of central portion of right female breast C50.112 malignant neoplasm of central portion of left female breast
- 34. Summary
- 35. Links Ef鍖cient Estimation of Word Representation in Vector Space (Mikolov) Distributed representation of words and phrases and their compositionality (Mikolov) word2vec Parameter Learning Explaining (Rong)
- 36. Questions?