Natural Language Processing
The NLP Service provides education and NLP consultation services to the University of Utah community and nationally, to improve the application of NLP to clinical and health-related problems.
1) Formal and informal education about clinical NLP - a journal club focused on NLP, an online course on information extraction, a data science summer school covering NLP, tutorials and lectures about NLP, and one-on-one consultation.
2) Self-serve access to NLP tools and data for development and research - hosting of websites to assist people in finding NLP tools for particular tasks (NLPToolFinder) and a collaborative environment attracting participation of a significant number of clinical NLP developers and researchers (NLP Ecosystem).
3) Contracted services for specific research and operational projects - consultation and contracting to collaborate on projects to automatically classify text, create summaries, and extract information. Our collaborative projects support both research and operations alike.
Examples of Operational/research projects in the NLP Service Line are:
- Interface for Word Vocabulary Expansion using Word Embeddings & SNOMED-CT
- Rule-based Segmentation Toolkit for extremely accurate Sentence Segmentation of Cinical Text
- Infrastructure for Reproducible Exposimic Research
- Improving Clinical Trial Cohort Definition Criteria with Distributed Semantic Matching
- Identifying surgical site infections from radiology reports
- Identifying carotid stenosis from carotid ultrasounds
- Classifying patient satisfation from pressganey.com survey comments
- Monitoring depression/anxiety through social media tweets
- Extracting measurement of colonoscopy exam quality
- Detecting critical findings rom radiology reports
- Extracting contraceptive methods from social histories
- Determining the duration of pain due to cholecystitis
- Classifying indications and reasons for medical exam
- Predicting the occurrence of cancer through family history data
Our collaborative projects support both research and operations. For more information on collaborating with or contributing to the NLP Service Line, please contact us.
What can NLP do for you?
- Extract information from clinical text, such as symptoms or procedures
- Summarize information in a clinical report.
- Identify topics in Twitter feeds.
- Classify patient comments as positive or negative
See our NLP projects page for more!