About

Jeff Clune is a pioneering researcher at the intersection of evolutionary computation, artificial intelligence, and robotics, whose work has fundamentally advanced how machines learn, adapt, and explore. Best known for his landmark 2015 paper "Robots that can adapt like animals" (948 citations), Clune demonstrated that robots could recover from damage by intelligently adapting their behavior — a breakthrough that brought biological resilience principles into robotics. His co-development of the MAP-Elites algorithm (412 citations) transformed how researchers search complex solution spaces, finding broad families of high-quality solutions rather than a single optimum, with applications spanning drug discovery to engineering design. Clune has made substantial contributions to reinforcement learning's hardest challenges, developing the Go-Explore framework (228 citations) and related approaches that tackle notoriously difficult sparse-reward environments like Montezuma's Revenge. His evolutionary AI research examines how complexity, hierarchy, and novelty-seeking emerge in both biological and artificial systems, including influential work on generative encodings for robot locomotion and the evolutionary origins of hierarchical networks. Across his career, Clune has consistently pushed the boundaries of open-ended learning, making him an indispensable figure for anyone studying adaptive AI systems.

Research Focus

Key Achievements

22
H-Index
31
Papers
3,385
Total Citations
109
Avg Citations/Paper
🏆 Most Cited Paper
Robots that can adapt like animals
948 citations · 2015
📈 Most Prolific Year: 2014 (7 Papers)
🤝 Key Collaborators: 40
🏛 Institutions: University of Wyoming, Uber AI (United States), Michigan State University, Cornell University, Wyoming Department of Education, University of British Columbia

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
    Unshackling evolution
    214 citations · 2013
  5. 5
  6. 6
  7. 7
  8. 8
    Unshackling evolution
    142 citations · 2014
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 13 days ago