Go-Explore paper out. New high/average scores: 18 million/650k on Montezuma's Revenge (44k w/o domain knowledge) & 100k+/~60k on Pitfall, all tested w/ sticky actions! Huge thanks to my co-authors Joost Huizinga Joel Lehman Jeff Clune Kenneth Stanley. arxiv.org/abs/1901.10995

Go-explore source code is now available! github.com/uber-research/… Paper: arxiv.org/abs/1901.10995 Blog: eng.uber.com/go-explore/ Joost Huizinga Joel Lehman Jeff Clune Kenneth Stanley github.com/uber-research/…

Newly accepted paper "Guiding Neuroevolution with Structural Objectives" (arxiv.org/abs/1902.04346) provides insights in how to leverage the power of structural organization in neural networks. Honored to have been a collaborator on this paper with Kai Olav Ellefsen and Jim Torresen!

Source code is now available for POET! It generates its own increasingly complex, diverse training environments & solves them, auto-generating curricula along the way. github.com/uber-research/… With Jeff Clune Joel Lehman & Kenneth Stanley Blog Uber Engineering : eng.uber.com/poet-open-ende…

Can we make an artificial brain more human? Should we? I recently spoke with Auburn University Professor of Computer Science Anh Nguyen (Totti) about his groundbreaking neural network research paving the way for future #AI journal.binarydistrict.com/ai-application…

We just released on arXiv our work on playing text adventure games wit Go-Explore in which we achieve surprising generalization results: arxiv.org/abs/2001.08868 Andrea Madotto Mahdi Namazifar Joost Huizinga Adrien Ecoffet Gokhan Tur Steven Zheng @Alexand90803163 Chandra Khatri

New paper! Scaling MAP-Elites to Deep Neuroevolution. We enable agents to recover from damage and explore in high-D control tasks, where traditional QD algorithms fail. Arxiv arxiv.org/abs/2003.01825. Work led by Cédric during his internship with me Uber AI Labs 1/3

Our paper on Guiding Neuroevolution with Structural Objectives (with Joost Huizinga and Jim Tørresen ) is out in this month's Evolutionary Computation Journal! mitpressjournals.org/toc/evco/28/1

Our work on playing text adventure games with Go-Explore has been accepted at #IJCAI2020 a big congrats to all the co-authors Mahdi Namazifar Piero Molino Joost Huizinga Adrien Ecoffet Steven Zheng @Alexand90803163 Dian Yu Chandra Khatri Gokhan Tur and Uber AI Labs

Thrilled to share "Multi-task curriculum learning in a complex, visual, hard-exploration domain: Minecraft"! Curriculum learning plus a dynamic exploration bonus enables agents to obtain items far up the Minecraft tech tree. Paper: arxiv.org/abs/2106.14876 youtu.be/MFDudOvn3oc