Artificial Intelligence —
Training activity: 6

Details

Develop a machine learning solution to a clinical problem

Type

Developmental training activity (DTA)

Evidence requirements

Evidence the activity has been undertaken by the trainee​.

Reflection on the activity at one or more time points after the event including learning from the activity and/or areas of the trainees practice for development.

An action plan to implement learning and/or to address skills or knowledge gaps identified.

Considerations

• Machine learning methods, including: supervised learning, unsupervised learning, semi-supervised learning and reinforced learning
• Regression, neural networks, random forest and nearest neighbour
• Fitting
• Bias

Reflective practice guidance

The guidance below is provided to support reflection at different time points, providing you with questions to aid you to reflect for this training activity. They are provided for guidance and should not be considered as a mandatory checklist. Trainees should not be expected to provide answers to each of the guidance questions listed.

Before action

• Consider the key stages involved in developing a machine learning solution, from problem definition to evaluation.
• What specific clinical problem will you be addressing? What data will be available? What kind of machine learning techniques might be suitable and why? How does addressing a clinical problem differ from addressing an administrative problem?
• Is the goal to explain or predict something? How does this influence your choice of technique?
• What specific insights do you hope to gain about the practical challenges and considerations of applying machine learning to real-world clinical problems?
• How will this activity develop your skills in the entire machine learning lifecycle, including problem formulation, data handling, model building, and evaluation?
• What is your current level of understanding and practical experience in developing machine learning solutions?
• Discuss potential clinical problems and available datasets with your training officer.
• Review relevant machine learning concepts, algorithms, and programming skills (e.g., Python).
• Consider the steps involved in developing a machine learning solution (e.g., data pre-processing, feature engineering, model selection, training, evaluation metrics).
• How do you feel about undertaking a practical machine learning development task?
• Do information governance or medical device regulations apply to your solution? How will that influence your approach to the task?

In action

• How are you approaching the development of the machine learning solution? What initial steps are you taking to define the problem and gather data?
• What decisions are you making about which algorithms or techniques to try as you start building the model?
• Which parts of the development process (e.g., data pre-processing, model training, evaluation) feel more familiar, and which require more focused attention and learning?
• How effectively are you addressing the clinical problem through your machine learning approach?
• What technical challenges are you encountering as you develop and test your solution? What clinical risks have you encountered and how have you mitigated them?
• What are you learning about the practical application of machine learning to clinical problems as the development unfolds?
• How does this development activity build upon your existing knowledge of programming, data analysis, and machine learning?
• If your initial approach to the problem or choice of algorithm is not yielding good results, what alternative strategies could you consider?
• Do you need to seek advice or resources to overcome specific technical hurdles you are facing?
• Are you ensuring that your development activities are complying with relevant regulations and ethical principles and within your current level of competency in machine learning?

On action

• Describe the clinical problem you addressed and the machine learning solution you developed. What were the key steps you took in developing your solution (e.g., data preparation, model selection, training, evaluation)? What challenges did you encounter during the development process?
• What practical skills in machine learning development did you gain or improve through this activity? Were there any unexpected difficulties or successes during the development process? What did you learn from these? How did your reflections-in-action (during development) influence the choices you made? How does this experience relate to applying machine learning techniques in your future professional practice?
• What specific aspects of machine learning solution development do you need to further refine? How will you apply the lessons learned from this project to future machine learning development tasks? What actions will you take to enhance your skills in developing machine learning solutions for healthcare problems? What support or resources might you need to further develop your machine learning development skills?

Beyond action

• Have you reviewed the machine learning solution you developed? How do you now view the problem you addressed and your solution in light of any new knowledge or experiences you have gained?
• Has the experience of developing a machine learning solution given you a deeper understanding of the challenges and complexities involved in applying AI to clinical problems? How has this training activity improved your problem-solving skills and your ability to apply computational thinking to healthcare challenges?
• What transferable skills (e.g., problem definition, data handling, algorithm development, evaluation) did you develop that will be valuable in future AI-related projects or research? What aspects of machine learning solution development (e.g., model deployment, clinical validation) would you like to learn more about?

Relevant learning outcomes

#	Outcome
# 4	Outcome Apply AI and machine learning techniques to address healthcare provision and clinical questions.