Personalizing fashion with human-in-the-loop machine intelligence

This is a highlight from a talk by Jianqiang Wang and Jasmine Nettiksimmons, “Combining statistics and expert human judgement for better recommendations.” Visit Safari to view the full session from the 2016 Artificial Intelligence Conference in New York.

While it seems that the definition of artificial intelligence is always evolving, it has remained fairly consistent that humans powered by machines outperform machines that work alone. Jay Wang and Jasmine Nettiksimmons discuss how this is true even as AI technologies make their way into helping consumers discover fashion and styles they love. Knowing where to place the human in the loop is critical and not easy, so Wang and Nettiksimmons give an inside look at how Stitch Fix systematizes collaboration between stylists and AI software to maximize their value creation together.

Related Resources:

• Humans in a loop: Jupyter notebooks as a front-end for AI pipelines at scale—session by Paco Nathan at JupyterCon, August 22-25, 2017, in New York City

• Active learning and transfer learning—session by Lukas Biewald at the O’Reilly Artificial Intelligence Conference in San Francisco, Sept. 17-20, 2017

• Artificial Intelligence: Teaching Machines to Think Like People, by Jack Clark

• “Building human-assisted AI applications“—Data Show Podcast with Adam Marcus