1. Chapter 12. Some Requirements for Human-Like Robots in Creating Brain-Like Intelligence, Aaron Solman. Course: Robots Learning from Humans Hur, Woo-Sol Vehicle Intelligence Laboratory School of Electrical and Computer Engineering Seoul National University http://vi.snu.ac.kr VEHICLE INTELLIGENCE LAB

2. Contents Introduction The Seduction of Embodiment Fallacies in Nouvelle AI Limitations of Symbolic AI Meta-semantic and Exosomatic Ontologies Morphology and Development 2

3. Contents Requirements for Human Visual Processing Two Extreme Kinds of Dynamical System Topics for Further Investigation The Variety of Possible Architectures Dynamical Systems for Symbolic AI Types of Architecture Layered Perception and Action How Evolution Produced Mathematicians? Conclusion 3

4. Some Questions  What animals did you see?  What was in the last picture?  What was in the picture taken at dusk?  Did you see any windows?  Did you see a uniformed official?  Did anything have horns? What?  Did anything have hands on the floor?

5. Requirements for Human Visual Processing  Human process photographs at a rate of one per second.  No known mechanism exists for human visual processing.  Reflection on a wide range of phenomena has led to a hypothesized architecture with a complex system.

6. Two Extreme Kinds of Dynamical System ■ 1 / 26  a simpler one  assumed to build “biologically inspired” robots  key feature is close coupling between internal states and the environment

7. Two Extreme Kinds of Dynamical System  more complex one

9. Topics for further investigation  how to build dynamical systems  how those systems are constructed over many years  how the constraint-propagation works  how the processes are influenced by various aspects  how the resulting percepts continue to be driven  how all of that is used

10. The Variety of Possible Architectures

11. Reminder Symbolic AI Symbol grounding Kant’s empiricism Learning of robots to discovery of statistical patterns relating sensory and motor signals Nouvelle AI (Brook) Emphasis embodiment and sensory-motor interactions with the environment Dispense with symbolic representations (But not fully rejected) Use morphology to reduce the SW sophistication

12. The Variety of Possible Architectures: Dynamical Systems for Symbolic AI  Existing symbolic AI systems are clearly nowhere near human competence except in very narrow domains:  chess (not Go), Mathematica, Matlab, compiler optimization, …  Symbolic AI will suffice for everything often fail to attend to the kinds of intelligence required for controlling continuous actions in a 3-D structured environment:  maintaining balance while pushing a broom, drawing a picture with a pencil, …

13. The Variety of Possible Architectures: Dynamical Systems for Symbolic AI  Nevertheless, it should be clear that the “traditional” sense→think/decide→act loop (presented in several AI textbooks) is much too restrictive to accommodate the requirements presented before.

14. The Variety of Possible Architectures: Types of Architecture  The CogAff schema

15. The Variety of Possible Architectures: Types of Architecture  The H-CogAff architecture

16. The Variety of Possible Architectures: Layered Perception and Action  Vision operates at different levels of abstraction:

17. How Evolution Produced Mathematicians?

19.  From Empirical to Mathematical Truths  turning around, counting objects, rubber band, …  Empirical discovery becomes a necessary truth.  It’s not a matter of probabilities.  different ways can be exist for human to discover useful affordances

20. Conclusion  Much of the work on embodied cognition in robots has focused on the terribly narrow problem of learning.  Symbolic computation was a human competence long before AI began, so there must be biological mechanisms that make it possible.

21. Conclusion  It will be necessary for AI researchers to abandon factional disputes and stop claiming of one key mechanism.  A synthesis of complementary approaches might be essential to progress in the long run

22. Q & A ■ 26 / 26

23. References  인공지능은 뇌를 닮아 가는가, 유신 저, 컬처룩, 2014  Aaron Sloman, “Some Requirements for Human-Like Robots: Why the Recent Over-Emphasis on Embodiment Has Held Up Progress”, Creating Brain-Like Intelligence, pp. 248-277, Springer (2009)  “Artificial Intelligence”, Wikipedia, Sep. 2015, https://en.wikipedia.org/wiki/Artificial_intelligence