Workshop / Overview
Artificial intelligence is becoming increasingly critical to develop innovative, competitive and differentiated medical businesses and products. Unfortunately, over 80% of AI projects fail, due to lack of proper vision, guidance, and data strategy.
Deployed AI solutions always contain a degree of human and societal biases that may influence the results and have costly consequences for organisations.
In this workshop, delivered by five AI and data scientists with strong medical and enterprise backgrounds, you will learn how to address the main challenges in building successful AI projects for clinical research, from identifying problems where AI can readily add value, to getting the right team and resources for your projects.
You will learn how to leverage human-centric design to mitigate risks in AI projects in digital medicine.
We will provide you compelling and intuitive examples of when, why and how things can go wrong at different stages of an AI project and what can be done about it. In the second part of this workshop, you will learn how to use containerized Docker environments on RENKU, an open source tool for integration of the data science workflows.
Together we will implement a solution for a medical problem in jupyter notebooks using neural networks with TensorFlow and Keras. You will see in practice how to build model performance KPIs that reconcile model accuracy, business and fairness metrics.
Workshop / Outcome
PART 1. MANAGERIAL SKILLS
- How to Innovate and differentiate through AI projects?
- What are the main challenges in AI projects?
- How to mitigate risks associated with AI?
- How to estimate the required resources (timeline, budget, team size)?
- Why do we need to address cognitive biases of users and how to do this?
PART 2. TECHNICAL SKILLS
- How to assess model impact on decisions?
- What is the right technology stack?
- How to build a reproducible and reliable workflow?
- What roles do I need for my AI project?
We offer 2h of free consulting with our team members on 3 selected, medical or bioinformatic related projects.
Workshop / Difficulty
Workshop / Prerequisites
Basic knowledge of Python and machine learning concepts. The participant must have his/her Laptop with internet access.
Track / Co-organizers
🚂 Flatland: Multi-Agent Reinforcement Learning for Trains
With Sharada Mohanty, Christian Scheller, Florian Laurent & Manuel Schneider09:00-17:00 September 28STCC, EPFL Campus, Lausanne, Switzerland
Federated Learning: collaborative machine learning on sensitive decentralized data
With Moritz Freidank, Aman Apte & Praveen Uggappakodi09:00-13:00 September 29STCC, EPFL Campus, Lausanne, Switzerland