Machine-learning algorithms are being used to help find patterns in data that people wouldn’t otherwise notice, and these are now being deployed to inform decisions in a variety of contexts.
However, machine learning is still a field for experts trained in the subject and it could be hard for novices to harness the power of machine learning for their requirements. But this is going to change as a new award-winning research from the Cornell Ann S. Bowers College of Computing and Information Science explores non-experts can leverage the power of machine learning effectively, efficiently and ethically to better enable industries beyond the computing field to harness the power of AI.
Swati Mishra, a Ph.D. student in the field of information science is lead author of “Designing Interactive Transfer Learning Tools for ML Non-Experts,” which received a Best Paper Award at the annual ACM CHI Virtual Conference on Human Factors in Computing Systems, held in May.
As artificial intelligence is expanding into new fields and industries, the need for research and effective tools to teach these skills to novice users is growing. Researchers have been working to bring machine learning to the masses but the study has generally been focused on understanding the users and the challenges they face when navigating the tools.
The new study including the development of new own interactive machine-learning platform breaks fresh ground by investigating the inverse: How to better design the system so that users with limited algorithmic expertise but vast domain expertise can learn to integrate preexisting models into their own work.
Mishra takes an unconventional approach with this research by turning to a complex process called “transfer learning” as a jumping-off point to initiate nonexperts into machine learning. Transfer learning is a high-level and powerful machine-learning technique typically reserved for experts, wherein users repurpose and tweak existing, pretrained machine-learning models for new tasks.
With this technique, you can use an existing model to identify new concepts without requiring extensive training.
Mishra’s research exposes transfer learning’s inner computational workings through an interactive platform so nonexperts can better understand how machines crunch datasets and make decisions. Through a corresponding lab study with people with no background in machine-learning development, Mishra was able to pinpoint precisely where beginners lost their way, what their rationales were for making certain tweaks to the model and what approaches were most successful or unsuccessful.