Development of an NLP Model for Clinical Note Transcription and Categorization
Project scope
Categories
Data visualization Data analysis Data modelling Databases Data scienceSkills
transcribing enthusiasm audio transcription natural language processing (nlp) data science python (programming language) application programming interface (api) machine learning gpt-3 (nlp model) generative artificial intelligenceWe are seeking enthusiastic data science and analytics students to work on an exciting project aimed at improving healthcare documentation through natural language processing (NLP). This project involves training an NLP model to transcribe audio recordings from clinical follow-up appointments and categorize the information into the SOAP (Subjective, Objective, Assessment, Plan) format. The project will utilize advanced machine learning techniques and cloud-based tools.
**Key Responsibilities**:
1. **Data Collection and Preparation**:
- Collect and preprocess audio recordings and transcriptions from clinical follow-up appointments.
- Annotate data with SOAP categories and generate synthetic data if necessary.
2. **Model Selection and Initial Training**:
- Choose and fine-tune a pre-trained NLP model (e.g., BERT, GPT-3).
- Implement transfer learning techniques and perform initial training on annotated datasets.
3. **Model Training and Validation**:
- Set up and execute training pipelines.
- Conduct hyperparameter tuning and validate model performance using appropriate metrics.
4. **API Deployment**:
- Deploy the trained model as an API endpoint.
- Develop API documentation and ensure seamless integration with other applications.
**
l
Key Responsibilities:
1. Data Collection and Preparation:
• Collect audio recordings and transcriptions from clinical follow-up appointments.
• Annotate the data with SOAP categories and create synthetic data where necessary.
2. Model Training and Validation:
• Select and fine-tune pre-trained NLP models using Vertex AI.
• Conduct hyperparameter tuning and validate model performance using appropriate metrics.
3. API Deployment:
• Deploy the trained model as an API endpoint using Vertex AI.
• Develop API documentation for seamless integration with other applications.
Required Skills:
• Experience in NLP and machine learning.
• Proficiency in Python and relevant ML libraries (TensorFlow, PyTorch, SpaCy).
• Knowledge of Google Cloud Platform, especially Vertex AI and Firebase.
Support Provided**:
1)**Technical Resources**:
- Access to Google Cloud Platform, including Vertex AI and Firebase.
- Cloud Skill Boost Innovator Program access for hands-on labs and learning.
- Access to Google's technical subscription program for resolving queries.
- Access to Google Meet for collaboration and meetings.
2 . **Training and Tutorials**:
- Training sessions on using Google Cloud services and other relevant tools.
- Access to tutorials and documentation for machine learning and NLP techniques.
3) **Feedback and Evaluation**:
- Regular feedback sessions to assess progress and provide constructive feedback.
- Evaluation of deliverables with detailed review and recommendations for improvement.
About the company
Mimico Physiotherapy and Chiropractic Clinic is a multidisciplinary health care facility serving the Mimico, Lakeshore-Parklawn and Queensway, South Etobicoke and Toronto communities.
Our team of physiotherapists, chiropractors and massage therapists practice a patient centered, evidence based comprehensive care model. Among the service offered are concussion treatment, pain management, acupuncture, pelvic floor therapy, vestibular rehab, massage including deep tissue massage, swedish massage, sports massage and more.
Our team is passionate about their work and driven to collaborate as a team to optimize your recovery often communicating with your physician, nurse practitioner and specialists to keep them informed of your recovery.