cacti ml
about
🌵 is a space for my (@samrawal) clinical AI work --My overall goals are to:
- Build Clinical AI software which
- Makes information deep within medical records more accessible: semantically filter, search, and extract clinical information from medical data like Electronic Medical Records
- Is generalizable: can be applied across a range of clinical tasks without significant task-specific training
- Develop packages, models, and tools that can be used as building blocks for further downstream clinical research and development
work
github.com/cactiML- My other non-clinical projects and research papers can be found on my personal website.
| software
- MedQA: a a GPT-powered clinical reference tool that can answer clinical questions and follow-up questions in natural language, along with references to sources. [More information]
- clinisift: multitool for processing clinical medical records.
- clinitokenizer: Sentence tokenizer for clinical/medical text.
| models
- bert-Clinical-NER: A Named Entity Recognition model for clinical entities (problem, treatment, test).
- bert-Med-NER: A Named Entity Recognition model for medication entities (medication name, dosage, duration, frequency, reason).
| datasets and resources
- Clinical GPT-3: Exploration of OpenAI GPT-3's abilities to perform various Natural Language Processing tasks in the clinical domain, on patient medical records.
- Sample Medical Notes: A collection of generated medical notes that can be used publicly for demos, etc. Coming soon.
contact
-
Sam Rawal: MD student @ University of Illinois Urbana-Champaign. Previously BS & MS in Computer Science @ Arizona State University.
Working on clinical NLP. 📚🥼🤖
samrawal.com | twitter.com/samarthrawal | github.com/samrawal