Algorithms that continually invent both problems and solutions

The Data Exchange Podcast: Kenneth Stanley on his work at Uber AI and starting a new group focused on open-endedness at OpenAI.

SubscribeiTunesAndroidSpotifyStitcherGoogle, and RSS.

In this episode of the Data Exchange I speak with Kenneth Stanley, a Senior Research Manager at Uber AI and a Professor at UCF. Ken just announced that starting in June he is starting a new research group focused on open-endedness at OpenAI.  He is a pioneer in the field of neuroevolution – a method for evolving and learning neural networks through evolutionary algorithms. Ken and his colleague, Joel Lehman, wrote one of my favorite books on AI aimed at a broad audience: Why Greatness Cannot Be Planned. In this episode we discuss his upcoming move to OpenAI, as well as his recent work on open-ended algorithms.

Scalable machine learning, scalable Python, for everyone: Join Michael Jordan, Manuela Veloso, Azalia Mirhoseini, Zoubin Ghahramani, Wes McKinney, Ion Stoica, Gaël Varoquaux, Raluca Popa and many other speakers at the first Ray Summit, a FREE virtual conference which takes place Sep 30th and Oct 1st.

Our conversation covered:

  • Ken’s new position at OpenAI.
  • The transition from being a longtime academic researcher to founding and helping lead an industrial research team (Uber AI Labs).
  • Open-ended algorithms, specifically his work on POET (Paired Open-Ended Trailblazer) and Enhanced POET.
  • Generative Teaching Networks.

Related content:

Subscribe to our Newsletter:
We also publish a popular newsletter where we share highlights from recent episodes, trends in AI / machine learning / data, and a collection of recommendations.

Back when I was still at O’Reilly, I recruited Ken to write these two seminal posts:

[Image: Abandoned Fish Processing Pier, Western End of Drakes Beach HDR 1 by Dean Wampler, used with permission.]