Kathryn Merrick

About

Kathryn Merrick is a prominent researcher whose work spans reinforcement learning, cognitive robotics, and intrinsic motivation in artificial systems. Her most influential contribution, "Hierarchical Deep Reinforcement Learning for Continuous Action Control" (2018, 197 citations), introduced a groundbreaking algorithm enabling robots to learn both basic and compound skills for complex tasks — a landmark advance in robotic control. Building on this, her 2017 work on multi-task deep reinforcement learning demonstrated remarkable efficiency gains, reducing network parameters by over 75% per task compared to conventional approaches. A defining thread throughout Merrick's career is her investigation of intrinsic motivation and value systems in autonomous agents. From early work exploring self-motivated exploration in robots (2010, 56 citations) to theoretical frameworks grounded in game theory and evolutionary multi-agent simulations, she has consistently pushed the boundaries of how artificial systems develop goal-directed behavior without explicit external reward. Her surveys on developmental cognitive robotics and trusted autonomy further establish her as a synthesizer of complex, interdisciplinary ideas. With a body of work accumulating over 440 citations across diverse venues, Merrick's research is essential reading for anyone interested in building robots and autonomous agents capable of genuinely adaptive, self-directed learning.

Research Focus

Key Achievements

Top Papers

1. 1
   Hierarchical Deep Reinforcement Learning for Continuous Action Control

   197 citations · 2018
2. 2
   Multi-Task Deep Reinforcement Learning for Continuous Action Control

   59 citations · 2017
3. 3
   A Comparative Study of Value Systems for Self-Motivated Exploration and Learning by Robots

   56 citations · 2010
4. 4
   Value systems for developmental cognitive robotics: A survey

   36 citations · 2016
5. 5
   A Review of Theoretical and Practical Challenges of Trusted Autonomy in Big Data

   35 citations · 2016
6. 6
   A game theoretic framework for incentive-based models of intrinsic motivation in artificial systems

   19 citations · 2013
7. 7
   Modeling Behavior Cycles as a Value System for Developmental Robots

   11 citations · 2010
8. 8
   Evolution of Intrinsic Motives in Multi-agent Simulations

   10 citations · 2012
9. 9
   Intrinsic Rewards for Maintenance, Approach, Avoidance, and Achievement Goal Types

   9 citations · 2018
10. 10
    Designing Toys That Come Alive: Curious Robots for Creative Play

    9 citations · 2008

Key Collaborators

Contact & Links