About

Riad Akrour is a leading researcher at the intersection of reinforcement learning (RL) and robotics, with a particular focus on making robot learning more sample-efficient, interactive, and human-centric. His foundational work on **preference-based reinforcement learning** has been highly influential—his 2012 paper “APRIL: Active Preference Learning-Based Reinforcement Learning” (94 citations) pioneered a paradigm where robots learn from human preferences rather than hand-crafted reward functions, dramatically reducing the burden on experts. This line of work, including “Preference-Based Policy Learning” (83 citations), established Akrour as a key figure in interactive robot education. Akrour has also made significant contributions to **hierarchical reinforcement learning** and **state abstraction**, developing methods that enable robots to learn complex, high-dimensional manipulation tasks. His 2018 paper on regularizing RL with state abstraction (21 citations) generalized abstraction to continuous action spaces, while his work on hierarchical tactile-based control (2020, 17 citations) advanced dexterous in-hand manipulation using tactile feedback. Across his career, Akrour’s research has consistently addressed the core challenge of bridging human intuition and machine learning, creating algorithms that are both theoretically sound and practically deployable on physical robots.

Research Focus

Key Achievements

8
H-Index
10
Papers
270
Total Citations
27
Avg Citations/Paper
🏆 Most Cited Paper
APRIL: Active Preference Learning-Based Reinforcement Learning
94 citations · 2012
📈 Most Prolific Year: 2012 (2 Papers)
🤝 Key Collaborators: 10
🏛 Institutions: Université Paris-Sud, Institut national de recherche en sciences et technologies du numérique, Technische Universität Darmstadt, Laboratoire de Recherche en Informatique, Fraunhofer Institute for Mechatronic Systems Design

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
    Interactive Robot Education
    9 citations · 2013
  9. 9
    Empowered skills
    4 citations · 2017
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 13 days ago