Open Science: Building Models LIke We Build Open-Source Software
- Liffey Hall 1
- 13:35 on 14 July 2022
- 30 minutes
The use of transfer learning has begun a golden era in applications of Machine Learning but the development of these models “democratically” is still in the dark ages compared to best practices in Software Engineering. I describe how methods of open-source software development can allow models to be built by a distributed community of researchers.
TalkPyData: Ethics in AI
Here, I elaborate on why we should develop tools that will allow us to build pre-trained models in the same way that we build open-source software. Specifically, models should be developed by a large community of stakeholders who continually update and improve them. Realizing this goal will require porting many ideas from open-source software development to building and training models, which motivates many threads of interesting research and opens up machine learning research for much larger participation.