TRAIL Research
The Responsible AI Learning Journey for Machine Learning Researchers
Description
As the architects of tomorrow’s technological landscape, ML researchers bear a critical responsibility in shaping our shared future. It is essential to equip researchers not only with the technical knowledge necessary to advance the frontiers of artificial intelligence, but also to help them develop an acute awareness of the broader societal implications of this technology, and help them assess the downstream impacts of the AI systems they develop.
Designed at Mila for our research community, the Trustworthy and Responsible AI Learning (TRAIL) certificate is a 12-hour synchronous training program for ML researchers. It provides participants with the foundational knowledge, skills, and tools necessary to responsibly design and conduct their AI research projects. Attendees learn to apply best practices and methodologies throughout the research cycle, from identifying ethical concerns to assessing impacts and investigating and mitigating unintended consequences.
Who Is It For?
This program is currently offered to Mila-affiliated students and researchers only.
However, if you are an educator interested in offering TRAIL to your students, our comprehensive Open Source Guide for educators will soon be available. With this free tool, you will have the content and facilitation notes required to offer the program to your students.
Open Source Guide for Educators: Interest Form
To receive the Guide as soon as it becomes available, please fill in this form.
For any additional information, you may contact ariana.seferiades@mila.quebec.
Learning Objectives
Recognizing the challenges researchers face in operationalizing responsible AI, TRAIL was created to equip Mila’s diverse pool of researchers (working on fundamental ML research to applied ML) with the essential tools and questions to practically implement RAI in their research projects.
Leveraging the vast knowledge of Mila-affiliated professors who are making significant contributions in the fields of FATE (Fairness, Accountability, Transparency, Ethics), human-AI interaction and responsible AI, this program was developed with the support and guidance of expert educators and researchers.
After this program, participants will be able to:
- Define what responsible AI and AI ethics are, and how it applies to the context of research;
- Develop ethical sensitivity and critical thinking skills to conduct responsible AI research;
- Learn to assess the downstream impacts of AI research projects and to make conscious design choices, through practical tools, frameworks and case studies;
- Learn how to adopt the AI research cycle as a strategic framework to consciously plan and execute responsible AI research projects;
- Making informed design decisions from the early stages of a project;
- Writing impact statements;
- Discover and apply socio-technical tools and best practices mitigate potential risks of AI systems.
This program exposes researchers to important concepts, ideas, methods and practices for pursuing AI research that is sensitive to downstream impacts and use cases. It helps to bring discussions around values and goals to machine learning.
Professor Jackie Cheung, Core Academic Member, Mila
93%
Satisfaction
93% of the participants would recommend the program to peer (100% on the last two cohorts)
120
Certifications
120 certified ML researchers since 2023
74%
Impact
74% of participants shared that the program significantly influenced their approach to ML research
Past Intructors
TRAIL covered a lot of topics surrounding responsible AI while being super fun and engaging! I especially loved the group workshop activities where we got to not only put into practice the concepts/tools we learned about, but also to socialize with people [...] Overall, it was an amazing experience that I recommend to all AI researchers [...].
Yu Lu Liu, MSc. Student, Mila/McGill University
Contact
If you have questions about the program, please contact Ariana Seferiades Prece, Project Manager, Learning at Mila.
