AV视频

Centre for Cognitive Science (COGS)

Spring term 2010

  • Week 1 (Jan 12th) Prof. Andrew Brown (Music, Queensland University of Technology, Australia): Generation in Context: Musical Enquiry through Algorithmic Music Making
  • Week 2 (Jan 19th) Prof Paul Coates (Hertfordshire): The Multiple Contents of Experience: Representation and Phenomenal Qualities
  • Week 3 (Jan 26th) Prof Mark Coeckelbergh (Twente, Holland): What is it like to be a robot? Natural empathy, artificial companions, and vulnerability mirroring
  • Week 4 (Feb 2nd) Dr Mark Bishop (Goldsmiths College, University of London): An introduction to Stochastic Diffusion Processes
  • Week 5 (Feb 9th) Dr Nicola Yuill (Sussex) and Prof Yvonne Rogers (Open University): Surface tension: Designing new interfaces to support or encourage?
  • Week 6 (Feb 16th) Prof Ezequiel Di Paolo (University of Basque Country, San Sebastian): The mind in-between: Can social interaction constitute social cognition?
  • Week 7 (Feb 23rd) No meeting
  • Week 8 (Mar 2nd) Prof John Pickering (Warwick): Artificial Life, Ethics and the Question of Limits
  • Week 9 (Mar 9th) Prof Luciano Floridi (Philosophy, Hertfordshire): Bodies of Information - e-Health and its Philosophical Implications
  • Week 10 (Mar 16th) Prof George F R Ellis (Mathematics, University of Cape Town): On the Nature of Top Down Causation in the Human Brain
  • Week 11 (Mar 23rd) Prof John Stewart (Compiegne): From autonomy to heteronomy (and back): The enaction of social life.


Week 1

Date: Tuesday, 12th January 2010
Speaker:
 Associate Professor Andrew R. Brown (Music, Queensland University of Technology, Australia)

Generation in Context: Musical Enquiry through Algorithmic Music Making

In this seminar I will discusses the use of generative music practice as a method for research. We have formalised an aspect of this approach we call, Generation in Context (GIC), designed to interrogate theories of music analysis and music perception. The GIC method involves building a computational model of a music-analytical or music-perceptual theory, and inverting the model (from analytical to generative) to create new musical works. The characteristics of these new works can be assessed against outcomes predicted by the theories.

This approach seeks to support the understanding of music by bringing together the study of music perception and algorithmic composition. Our objectives are to: (i) Use algorithmic composition systems to interrogate models of music perception or music theory, and (ii) Develop improved algorithmic compositional techniques by incorporating models of music perception.

I will outline some explorations using generative processes for music research, and will discuss how generative processes provide evidence as to the veracity of theories about how music is experienced, with insights into how these theories may be improved and, concurrently, provide new techniques for music creation.


Week 2

Date: Tuesday, 19th January 2010
Speaker:
 Prof Paul Coates (Hertfordshire)

The Multiple Contents of Experience: Representation and Phenomenal Qualities

Perceptual, and other kinds of experiences such as hallucinations, have two important features. They represent states of affairs, being directed onto external physical objects. They also have a distinctive phenomenology, involving the awareness of phenomenal qualities, which accounts for the distinctive sensory character of perception. This paper explores the relationship between these features, and focuses on the views of those Intentionalists who attempt to argue for a form of Direct Realism about perception. Motivated by assumptions about the supposed transparency of perception, Intentionalists such as Michael Tye claim that the qualities of which we are aware in veridical perceptual experience are qualities belonging to external things. I argue that there is no satisfactory analysis of the notion of representation which can support this version of Direct Realism.

I sketch out reasons for accepting a two-component view of perceptual experience: We should recognise that phenomenal qualities belong to inner states, which are caused by the objects we perceive, and which in turn cause and guide intentional states (dubbed by Sellars, "perceptual takings") focused directly upon external objects. These considerations support the view that a given token perceptual experience has (at least) three distinct forms of content. Two of these derive from the role that perception plays in enabling subjects to navigate through the local environment, in order to satisfy their needs.

Selected references:
Michael Tye, Consciousness, Color, and Content, MIT 2000; Consciousness Revisited, MIT 2009;

  • Paul Coates, The Metaphysics of Perception, Routledge 2007; 

  • Susanna Siegel, 'The Contents of Perception', Stanford Encyclopedia of Philosophy; 

  • Wilfrid Sellars, 'The Role of the Imagination in Kant's Theory of Experience', reprinted in: In the Space of Reasons, (eds) Kevin Scharp and Robert Brandom, Harvard UP 2007.

Week 3

Date: Tuesday, 26th January 2010
Speaker:
 Prof Mark Coeckelbergh (Twente, Holland)

What is it like to be a robot? Natural empathy, artificial companions, and vulnerability mirroring

Can robots become companions, and if so, what are the necessary conditions and what are the ethical issues that might arise in human-robot companionship relations? I argue that the possibility and future of robots as companions depends (among other things) on the robot's capacity to be a recipient of human empathy and that a necessary condition for this to happen is that the robot mirrors human vulnerabilities. For the purpose of these arguments, I make a distinction between empathy-as-cognition and empathy-as-feeling, connecting the latter to the moral sentiment tradition and its concept of 'fellow feeling'. Furthermore, I show that we might have the intuition that vulnerability mirroring raises the ethical issue of deception, but that given contemporary technological developments we cannot justify the underlying assumption that artificial vulnerability is less real (and less significant or valuable) than natural vulnerability. On reflection, it becomes increasingly difficult to maintain a strict conceptual distinction between artificial and natural. I conclude that if we want to hold on to the deception argument, we need a convincing answer to these objections.


Week 4

Date: Tuesday, 2nd Febuary 2010
Speaker:
 Dr Mark Bishop (Goldsmiths College, University of London)

An introduction to Stochastic Diffusion Processes

Stochastic Diffusion Processes (SDPs) define a class of Swarm Intelligence algorithms that can be simply applied to some problems in search and optimisation. The algorithmic processes behind these algorithms have been extensively mathematically described, hence formal algorithmic behaviour (convergence; time complexity etc) under a wide variety of situations has been established. SDPs are potentially interesting to a variety of researchers in A.I. and Cognitive Science as (i) they are a family of global search/optimisation/resource-allocation algorithms that have been usefully applied to a number of problems - ranging from wireless telephony (site selection) to robot localisation - for which their algorithmic behaviour is well understood; (ii) their behaviour mirrors a recruitment process observed in one species of ant; (iii) they can be easily instantiated on a connectionist architecture; (iv) they describe a mode of processing that naturally manipulates 'knowledge' greater than arity (0); (v) at convergence, semantic knowledge (of the task solution) is not [spatially] bound to any one agent, but is maintained by a [statistically] stable set of outputs across an ever changing population of agents.

This talk will present a simple metaphor that describes the operation of SDPs and review results from papers which established its core computational and algorithmic properties.


Week 5

Date: Tuesday, 9th Febuary 2010
Speaker:
 Dr Nicola Yuill (Sussex) and Prof Yvonne Rogers (OU)

Surface tension: Designing new interfaces to support or encourage?

An exciting new technological development that is on the brink of invading our classrooms is the interactive tabletop, such as Microsoft's Surface or the SmartTable. But will the introduction of these surfaces live up to the hype, or will they go the same way as the whiteboard - a handy networked powerpoint presentation surface, but not a revolution in collaboration? In this talk we will describe studies from the ShareIT project () to illustrate how digital surfaces can provide the conditions for good collaboration but cannot force people to work together. We consider psychological underpinnings - awareness of the other, joint attention and joint action - and how these need to be part of our investigation of technology-supported social activity.


Week 6

Date: Tuesday, 16th Febuary 2010
Speaker:
 Prof Ezequiel Di Paolo (University of Basque Country, San Sebastian)

The mind in-between: Can social interaction constitute social cognition?

Recent empirical work in social cognition, both in psychology and neuroscience, has gradually started to focus on situations involving various degrees of social interaction. Such situations are notably difficult to manage in controlled settings. This is one reason to account for the prevailing attention to individual cognitive mechanisms for social understanding. However, the "experimental quarantine" (Daniel Richardson's phrase) is being lifted and the focus of empirical studies is increasingly concerned with individuals in interactive situations (e.g., joint action).

In this talk, I argue that this move must be followed by a lifting of the "conceptual quarantine" still in effect, which puts the weight of social cognitive performance solely on individual mechanisms. This perspective is traceable to the methodological individualism prevalent in cognitive science in general. It is yet another reason to account for the widespread conception of social cognition as a detached observation of social situations and exceptionally as a form of participation. The properties of the interaction dynamics are relegated to the role of informational input to individual mechanisms.

In order to conceive of the possibility of social interaction being itself part of the mechanisms of social cognition, it is necessary first to provide a definition of the term able to capture the intuitive notion of engagement. Such definitions are surprisingly rare in the literature. I argue that the enactive definition of social interaction achieves this objective. Following this, the possible roles that interaction could play in particular cases are evaluated according to a scale of increasing involvement by introducing distinctions between contextual factors, enabling conditions and constitutive processes. I discuss existence proofs for all of these options (thus answering the title question positively).

The argument carries minimal and maximal implications. At the very least, if the interaction process is admitted to play in some cases a role beyond the contextual, this implies that individual mechanisms (e.g., contingency detection modules, mirror neurons) must be re-conceptualised as mechanisms-in-interaction, and their functional role re-assessed. I discuss evidence that this shift is slowly taking place for the case of mirror neurons. Maximally, if interaction is admitted to sometimes constitute social cognition this opens the door for a broadening of the spectrum of explanations and calls for a program aimed at assessing the contributions of individual and social mechanisms not only for social cognition, but for cognition in general.


Week 7

Date: Tuesday, 23rd Febuary 2010
No meeting


Week 8

Date: Tuesday, 2nd March 2010
Speaker:
 Prof John Pickering (Warwick)

Artificial Life, Ethics and the Question of Limits

From La Mettrie to Transhumanism, the effort to understand human existence in scientific terms has gathered momentum. With the rise of Artificial Life, the mechanisation of life and mind has taken a new and powerful turn, making it easier than ever to treat organisms as machines and machines as organisms. This has the potential to unleash enormous social and political forces. When cultures cannot contain such forces, the outcome is violence, as Walter Benjamin recognised.

Attempts to transcend human biological and psychological limitations belong in an honourable, if occasionally dangerous, tradition containing Daedalus and Babel as well as algebra and antibiotics. This talk will sketch some of the possible ethical issues raised by projects such as transhumanism, cognitive enhancement and sociable technology.

About the speaker:

John Pickering's present focus is on bringing together Merleau-Ponty, biosemiotics and embodied cognition. He sees this as contributing to the re-emergence of a process worldview, compatible with those of Alfred North Whitehead and C.S. Peirce.

John has an enduring interest in AI, Artificial Life and the integration of ICT into human society especially during development.

He is also interested in the potential for combining Western psychology and Eastern traditions, especially Buddhism. He has worked in a number of countries with Buddhist traditions including Sri Lanka, India, Korea, Thailand and China.


Week 9

Date: Tuesday, 9th March 2010
Speaker:
 Prof Luciano Floridi (Philosophy, Hertfordshire and Oxford)

Bodies of Information - e-Health and its Philosophical Implications

The first part of the talk will introduce an interpretation of the information turn as a fourth revolution. We are not immobile, at the centre of the universe (Copernican revolution); we are not unnaturally detached and diverse from the rest of the animal world (Darwinian revolution); we are not Cartesian subjects entirely transparent to ourselves (Freudian). We are now coming to see that we are not disconnected entities, but rather informational organisms, sharing with biological agents and engineered artefacts a global environment ultimately made of information, the infosphere (Turing revolution). In the second part, the previous framework will be used to understand the development of e-Health and its ethical issues. The fourth revolution is increasingly affecting our views about human nature, its fragility and resilience, its health (including mental health) and how we may shape it and make it flourish. We shall see how human bodies may be interpreted informationally and what this will mean, in the future, in terms of their well-being.


Week 10

Date: Tuesday, 16th March 2010
Speaker:
 Prof George F R Ellis (Mathematics, University of Cape Town)

On the Nature of Top Down Causation in the Human Brain

Abstract: The concept of top-down causation is important in understanding the human brain. I suggest that five different classes of top-down causation occur at different scales. They are, 
1. algorithmic top-down causation; 
2. top-down causation via non-adaptive information control; 
3. top-down causation via adaptive selection; 
4. top-down causation via adaptive information control; and 
5. intelligent top-down causation (i.e. the effect of the human mind on the physical world). 
Because of the existence of random processes at the bottom, there is sufficient causal slack at the physical level to allow all these kinds of causation to occur without violation of physical causation. That they do indeed occur is indicated by many kinds of evidence.


Week 11

Date: Tuesday, 23rd March 2010
Speaker:
 Prof John Stewart (Compiegne)

From autonomy to heteronomy (and back): The enaction of social life.

The term "social cognition" can be construed in different ways. On the one hand, it can refer to the cognitive faculties involved in social activities, defined simply as situations where two or more individuals interact. On this view, social systems would consist of interactions between autonomous individuals; these interactions form higher-level autonomous domains not reducible to individual actions. A contrasting, alternative view is based on a much stronger theoretical definition of a truly social domain, which is always defined by a set of structural norms; moreover, these social structures are not only a set of constraints, but actually constitute the possibility of enacting worlds that would just not exist without them. This view emphasises the heteronomy of individuals who abide by norms that are impersonal, culturally inherited and to a large extent independent of the individuals. Human beings are socialised through and through; consequently, all human cognition is social cognition. The article argues for this second position. Finally, it appears that fully blown autonomy actually requires heteronomy. It is the acceptance of the constraints of social structures that enables individuals to enter new realms of common meaningfulness. The emergence of social life marks a crucial step in the evolution of cognition; so that at some evolutionary point human cognition cannot but be social cognition.

Keywords Social cognition - Autonomy - Heteronomy - Normativity