Alejandro Agostini

About

Alejandro Agostini is a prominent researcher in cognitive robotics and artificial intelligence, whose work centers on enabling robots to operate intelligently and adaptively within human environments. He is best known for developing the theoretical and computational framework of **Object-Action Complexes (OACs)**, grounded representations that link perception and motor processes to support robotic cognition — a concept that has garnered over 150 citations and become foundational in the field. His early contributions established procedural cognitive architectures built around OACs, shaping how autonomous agents model and predict sensorimotor interactions. Agostini's research has consistently tackled the challenge of making robots learnable by non-expert users, integrating AI planning with interactive machine learning so robots can rapidly acquire new behaviors without disrupting ongoing tasks. His work on object substitution — allowing robots to recognize functional alternatives for missing tools, much as humans might use a knife as a screwdriver — demonstrates a creativity that bridges symbolic reasoning and real-world manipulation. More recently, he has extended this vision through functional object-oriented networks for long-horizon planning and efficient state abstraction. With publications spanning over a decade and accumulating hundreds of citations, Agostini's contributions represent a sustained and impactful effort to bring genuinely intelligent, human-compatible robots closer to reality.

Research Focus

Key Achievements

Top Papers

1. 1
   Object–Action Complexes: Grounded abstractions of sensory–motor processes

   153 citations · 2011
2. 2
   Cognitive agents — a procedural perspective relying on the predictability of Object-Action-Complexes (OACs)

   75 citations · 2008
3. 3
   Efficient interactive decision-making framework for robotic applications

   47 citations · 2015
4. 4
   Integrating task planning and interactive learning for robots to work in human environments

   29 citations · 2014
5. 5
   A cognitive architecture for automatic gardening

   21 citations · 2017
6. 6
   Using structural bootstrapping for object substitution in robotic executions of human-like manipulation tasks

   15 citations · 2015
7. 7
   Long-Horizon Planning and Execution With Functional Object-Oriented Networks

   11 citations · 2023
8. 8
   Simultaneously learning at different levels of abstraction

   9 citations · 2015
9. 9
   Action rule induction from cause-effect pairs learned through robot-teacher interaction

   8 citations · 2008
10. 10
    Efficient State Abstraction using Object-centered Predicates for Manipulation Planning

    4 citations · 2020

Key Collaborators

Contact & Links