[TMP-072] LWMs teaching to teach
An LWM (large whatever model) tailored to nurse training
High-quality education and training for nurses are essential to maintaining high medical care standards. However, as highlighted by the COVID-19 pandemic, there is a critical shortage of healthcare professionals. Accelerating nurse training is necessary to meet demand, but this often risks reducing quality, potentially leading to poor qualifications or harmful outcomes. A key challenge is optimizing training to accelerate learning while maintaining quality.
Teachers must assess each student’s educational progress: who needs more practice, who can advance, and who is ready to work with patients. Effective training optimization involves personalizing both student learning and the feedback teachers receive about their instruction. This is where Artificial Intelligence (AI), particularly foundational models like large language models (LLMs) paired with other machine learning techniques, can provide valuable support.
In collaboration with nurse educators from the University of Southampton, this microproject will design an LWM (large whatever model) tailored to nurse training. Nurse teachers will visit DFKI to explore available systems, while DFKI researchers will observe training in Southampton. The defined LWM will be tested on recorded training sessions, focusing on understanding trainees' actions, identifying errors, evaluating teaching strategies, and providing actionable feedback for both students and educators.
Building models of medical procedures require efforts that go beyond the scope and time frame of a micro-project. Therefore, this work is still ongoing and will proceed after the end of the Humane AI Net.
So in regards of project result at the time Humane AI Net ended is:
- identification of scenarios with a potential for generative AI to benefit health training – training of cannulation and venipuncture
- defining a procedure how to introduce Generative AI in training of cannulation and venipuncture.
- planning a study towards developing the required LWM models
- recording an extensive data-set in an actual medical training facility following actual training procedures.
- starting the long process of data processing and algorithm development (which is ongoing)
We collected a dataset consisting of: 90h of video (20 person recording 4 sessions of about 20+ min each, from 3 different cameras) accompanied with respective IMU Data + GoPro user view + audio recording and expert feedback of the process of cannulation and venipuncture.
Tangible Outcomes
- [arxiv] Stefan Fritsch and Matthias Tschoepe and Vitor Fortes Rey and Lars Krupp and Agnes Gruenerbl and Eloise Monger and Sarah Travenna, GenAI Assisting Medical Training, arXiv, mobiCHAI workshop in MobileHCI2024 https://arxiv.org/abs/2410.16164
- presented at: mobiCHAI – 1st International Workshop on Mobile Cognition-Altering Technologies (CAT) using Human-Centered AI, at The ACM International Conference on Mobile Human-Computer Interaction Melbourne, Australia https://ai-enhanced-cognition.com/mobichai/
Partners
- DFKI, EI, Agnes Grünerbl
- Health Department, University of Southampton, Eloise Monger