Fostering Multidisciplinary Collaboration in Artificial Intelligence and Machine Learning Education: Tutorial Based...

Area of Science:

• Biomedical research
• Artificial Intelligence
• Machine Learning

Background:

• Limited interdisciplinary training exists for AI/ML in biomedical research, creating a gap between data scientists and clinicians.
• The NIH's Bridge2AI initiative launched AI-READI to create a multimodal, FAIR dataset for diabetes research.
• AI-READI aims to train a workforce proficient in computational methods and clinical applications.

Purpose of the Study:

• To describe the design, refinement, and outcomes of the AI-READI Bootcamp.
• To share lessons learned for developing future multidisciplinary AI/ML training programs in biomedical research.

Main Methods:

• An 80-hour bootcamp combining lectures, coding, and mentorship, refined annually based on participant feedback.
• Year 1 focused on foundational Python and ML techniques; Year 2 integrated the AI-READI dataset and added modules on large language models and FAIR data principles.
• Participant satisfaction and characteristics were assessed via pre/post surveys, with qualitative feedback analyzed thematically.

Main Results:

• High participant satisfaction reported across both years, with Year 2 showing improvements due to smaller cohorts and applied learning.
• Year 2 achieved perfect scores for instructor effectiveness, staff support, and overall enjoyment.
• Participants valued working with multimodal biomedical datasets, peer collaboration, and the applicability of learned skills.

Conclusions:

• The AI-READI Bootcamp demonstrates a successful model for bridging technical and clinical expertise in biomedical AI through feedback-driven, multidisciplinary training.
• Key elements include diverse cohorts, applied learning with relevant datasets, and sustained mentorship.
• Future iterations will include pre-bootcamp modules, objective skill assessments, and long-term tracking of research productivity.