Reifying Robots

To reify is "to convert mentally into a thing", according to the OED.  What does it mean, computationally, for a robot to reify something?  What kinds of advantages and disadvantages do representations that reify offer the builders of robots?  Can we have first-order representations that admit uncertainty and still maintain moderately tractable planning, learning, and inference?  Can we extract such representations from the visual signal?

We'll try to understand these questions and more in a reading group. 

Everyone is welcome, independent of representation-religious preference.
 
 

Reading for Wednesday, December 15

Previous readings:

• September 22: Real-time control of attention and behavior in a logical framework,, I. Horswill.
• September 27: Pengi: An Implementation of a Theory of Activity, Philip E. Agre and David Chapman, AAAI87.
• October 6: Chapter 7 of Reinforcement Learning with Selective Perception and Hidden State, Andrew Kachites McCallum.
• October 13: Object-Oriented Bayesian Networks, D. Koller and A. Pfeffer
• October 20 and 27: Map Learning with Uninterpreted Sensors and Effectors, David Pierce and Benjamin Kuipers.
• November 3: "Object perception, object-directed action, and physical knowledge in infancy," Spelke, E.S., Vishton, P.M., & von Hofsten, C. (1994). In Gazzaniga, M. (Ed.), The Cognitive Neurosciences. Cambridge: MIT Press. Abstract.
• November 10: Towards Concept Formation Grounded on Perception and Action of a Mobile Robot,V. Klingspor and K. Morik

Possible future readings:

• Unifying Segmentation, Tracking, and Visual Search. I. Horswill and C. Barnhart
• Probabilistic frame-based systems, D. Koller and A. Pfeffer
• P-Classic: A tractable probabilistic description logic,  D. Koller, A. Levy, and A. Pfeffer
• Learning probabilities for noisy first-order rules, D. Koller and A. Pfeffer
• Inductive Logic Programming
•  Lifeworld analysis, P. Agre and I. Horswill
•  Growing Ontologies P. Cohen
•  Probabilistic Horn abduction and Bayesian networks , D. Poole
•  LP---A Logic for Representing and Reasoning with Statistical Knowledge , F. Bacchus
•  On the Origin of Objects, B.C. Smith
• Crangle & Suppes on robot language
• Image segmentation and object perception in humans (need good suggestions here)
• Inductive Logic Reinforcement Learning, Dzeroski & DeRaedt, ICML98