Wednesday, May 31, 2006

Evolving concepts of creativity: A mirror, a tightrope and an inkblot

A few hours ago I spoke at the second University of Sussex creativity workshop, "Evolving Views of Creativity". Speaking near the end of the day, my role was to synthesize what had been said, in aid of developing a consensus on what we at Sussex mean by creativity. For these reasons, this lecture will be of most interest to people at Sussex, but others may get something out of it as well.

My main point is that creativity can be seen as the result of maintaining a fruitful tension between:

  • Self and environment
  • The intuitive and the conceptual
  • "Blue-sky" and applied
  • Novel and familiar
  • Chaos and order
  • Disconnection and engagement.


Media:

Tuesday, May 23, 2006

Counterfactual computational vehicles of consciousness

Given April 7th 2006 in Tucson at Toward a Science of Consciousness 2006, this is really two talks in one. My attendance at the conference was made possible in part by grant OCG43170 from the British Academy; I am grateful for their support.

Abstract:
In a recent paper (Bishop 2002), Bishop argues against computational explanations of consciousness by confronting them with a dilemma. On a non-causal or weakly causal construal of computation, familiar arguments from Putnam and Searle reveal computation to be too observer-relative to be able to underwrite any law-involving explanation of consciousness. On a (much more plausible) strongly causal notion of computation (Chalmers 1994; Chalmers 1996; Chrisley 1994), computational explanations must advert to non-actual, counterfactual states and state transitions. Working this fact into versions of Fading Qualia and Suddenly Disappearing Qualia arguments, Bishop concludes that strongly causal computationalism cannot be physicalist, in that it maintains that two states may differ only their non-physical (i.e. counterfactual) properties and yet be phenomenally distinct. I rebut this argument by embracing the second horn, and denying that appeal to non-actual or counterfactual properties is at odds with physicalism; indeed, it is the lifeblood of normal, physical, causal explanation. I further cast doubt on the argument by showing that it is too strong; if it is correct, computational states could not explain anything at all, not even computational phenomena, let alone conscious experience. I show how computational states that differ in their counterfactual properties must, contra Bishop's characterization, differ with respect to some of their actual properties. However, I note that inter-dependencies between current experience and computational state may only be explicable by reference to explicit, counterfactual states rather than the occurrent physical states which realize those dispositional properties. This is shown to cohere with at least one understanding (Chrisley 2004) of the sensorimotor contingency theory of perceptual experience (O'Regan and Noe 2000), in which expectation is understood as a disposition to produce a computational state corresponding to the sensation one would have if one were to perform a particular action. I conclude by sketching some implications for the search for correlates of experience. The considerations arising out of Bishop's argument show that if computationalism is true, then the search for correlates will fail if it only considers occurrent non-dispositionally construed physical states at a time to be the possible correlates of the experience being had at that time.

References:

  • Bishop, J.M. (2002) "Dancing With Pixies", in Preston, J. & Bishop, J.M., (eds), Views into the Chinese Room, pp. 360-379, Oxford University Press.
  • Chalmers, D.J. (1994) "On Implementing a Computation", Minds and Machines, vol.4, pp.391-402.
  • Chalmers, D. (1996) "Does a Rock Implement Every Finite-State Automaton?", Synthese, vol.108, pp.309-333.
  • Chrisley, R. (1994) "Why Everything Doesn't Realize Every Computation," Minds and Machines 4:4, pp 403-420.
  • Chrisley (2004) "Perceptual Experience as the Mastery of Sensorimotor Representational Contingencies", abstract in Proceedings of Towards a Science of Consciousness 2004, p 119.
  • O'Regan, K., and Noe, A. (2001) "A Sensorimotor Account of Vision and Visual Consciousness", Behavioral And Brain Sciences 24(5).


Media:

In defense of transparent computationalism

This talk, given on May 5th 2006 in Laval, France at the International Conference on Computers and Philosophy, was originally to be based on a paper I wrote in 1999, but ended up diverging from it substantially.

Abstract of the original 1999 paper:

"A distinction is made between two senses of the claim “cognition is computation”. One sense, the opaque reading, takes computation to be whatever is described by our current computational theory and claims that cognition is best understood in terms of that theory. The transparent reading, which has its primary allegiance to the phenomenon of computation, rather than to any particular theory of it, is the claim that the best account of cognition will be given by whatever theory turns out to be the best account of the phenomenon of computation. The distinction is clarified and defended against charges of circularity and changing the subject. Several well-known objections to computationalism are then reviewed, and for each the question of whether the transparent reading of the computationalist claim can provide a response is considered."

I added to this by claiming that Gödel-style arguments don't show AI is impossible, but rather that the Church-Turing thesis is false. I rejected currently fashionable notions of pan-computationalism in favour of a view that makes having semantic properties essential to computation. I also argued that even if computationalism turns out to be false, it might still be possible for an artificial computational system to have a mind by virtue, at least in part, of the program it is running, since programming a computer not only changes it functionally, but also physically.

Media:

Monday, May 22, 2006

Finding aesthetic pleasure on the subjective edge of chaos: A proposal for robotic creativity

This is a lecture I gave at Goldsmiths College in London on May 16th 2006 as part of a Workshop on Computational Models of Creativity in the Arts.

In the talk, I give the nine axioms that constitute my approach to creating systems that exhibit creativity. The axioms are:


  • Axiom 1: If you make your robot pleasure-seeking, and make creativity pleasurable, you'll make your robot creative
  • Axiom 2: To be a good creator, it helps to be an appreciator
  • Axiom 3: Let the robot experience output in the real world, as we do
  • Axiom 4: We won’t like what it likes unless it likes what we like
  • Axiom 5: An important motivator is the approval or attention of others
  • Axiom 6: Sometimes it is better not to try pursue novelty directly, but something that is correlated with it
  • Axiom 7: Let dynamics play a role in appreciation
  • Axiom 8: Patterns in one's own states can be the objects of appreciation
  • Axiom 9: The best way to make outputs in the real world is to be embodied in the real world

But you'll have to listen to the talk if you want to know what all that means!

Media: