Papers
16
Total Citations
192
H-Index
7
About
Matthew Luciw is a leading researcher in developmental robotics and intrinsically motivated learning, with a focus on enabling autonomous agents to acquire skills through curiosity-driven exploration. His most influential work introduces Curiosity-Driven Modular Incremental Slow Feature Analysis (CD-MISFA), a groundbreaking framework that allows robots to autonomously build multiple stable sensory representations from high-dimensional video inputs. This approach, detailed in his highly cited 2015 paper (51 citations), enables humanoid robots to continually learn new skills without external rewards, simply by following their own curiosity. Luciw's research demonstrates that artificial curiosity—where an agent seeks to maximize its learning progress—can drive the acquisition of complex behaviors, from a Katana robot arm learning to manipulate wooden blocks (30 citations) to full skill acquisition systems like SKILLABILITY. His contributions extend to reinforcement learning with confidence-based goal generation and efficient interactive learning from binary feedback. With over 175 total citations across his top papers, Luciw has established himself as a pioneer in creating autonomous learning systems that develop abstractions and skills through intrinsic motivation, fundamentally advancing how robots can learn in open-ended environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Learning skills from play: Artificial curiosity on a Katana robot arm30 citations · 2012
- 4
- 5Explore to see, learn to perceive, get the actions for free: SKILLABILITY11 citations · 2014
- 6
- 7
- 8Efficient Interactive Multiclass Learning from Binary Feedback7 citations · 2014
- 9Autonomous reinforcement of behavioral sequences in neural dynamics7 citations · 2013
- 10