The Whitehead Lectures in "Cognition, Computation and Culture"
The Departments of Computing and Psychology at Goldsmiths organise regular seminars by guest speakers throughout the academic year encompassing various aspects of cognition, computation and culture. All are welcome to attend.
All seminars to be held at 4pm in the Pimlott Lecture Theatre, (Ben Pimlott Building), unless otherwise stated. (For directions to Goldsmiths see: www.goldsmiths.ac.uk/find-us).
To be added to the seminar mailing list, please contact Mark Bishop by email: m.bishop (@gold.ac.uk)
Lent 2010
Prof. Peter Robinson, Professor of Computer Technology, University of Cambridge Computer Laboratory, UK
Computation of emotion in man and machine
The importance of emotional expression as part of human
communication has been understood since the seventeenth century, and
has been explored scientifically since Charles Darwin and others in the
nineteenth century. Advances in computer technology now allow machines
to recognise and express emotions, paving the way for improved
human-computer and human-human communications. This talk presents some
recent advances in theories of emotion and affect, their embodiment in
computational systems, the implications for general communications, and
broader applications.
Recent advances in Psychology have greatly improved our understanding
of the role of affect in communication, perception, decision-making,
attention and memory. At the same time, advances in technology mean
that it is becoming possible for machines to sense, analyse and express
emotions. We can now consider how these advances relate to each other
and how they can be brought together to influence future research in
perception, attention, learning, memory, communication, decision-making
and other applications.
The computation of emotions includes both expression and recognition,
using channels such as facial expressions, non-verbal aspects of
speech, posture, gestures and general behaviour. The combination of new
results in psychology with new techniques of computation on new
technologies will enable new applications in commerce, education,
entertainment, security, therapy and everyday life. However, there are
important issues of privacy and personal expression that must also be
considered.
Brief Bio: Peter Robinson is Professor of Computer Technology in the
Computer Laboratory at the University of Cambridge, where he leads the
Rainbow Research Group working on computer graphics and interaction.
His research concerns problems at the boundary between people and
computers. This involves investigating new technologies to enhance
communication between computers and their users, and new applications
to exploit these technologies. The main focus for this is
human-computer interaction, where he has been leading work for some
years on the use of video and paper as part of the user interface. The
idea is to develop augmented environments in which everyday objects
acquire computational properties through user interfaces based on video
projection and digital cameras. Recent work has included desk-size
projected displays and inference of users' mental states from facial
expressions, speech, posture and gestures.
Professor Robinson is a Fellow of Gonville & Caius College where he
previously studied for a first degree in Mathematics and a PhD in
Computer Science under Neil Wiseman. He is a Chartered Engineer and a
Fellow of the British Computer Society.
27 January, 2010
Professor Hermann Müller, Professor of Experimental/Cognitive Psychology,
Universities of Munich, Germany, and London (Birkbeck College)
What pops out in pop-out?
A dimension-weighting account of visual search for salient pop-out targets.
Visual search for salient singleton targets, such as a red object amongst green (distractor) objects, is surprisingly efficient - phenomenally, such targets appear to 'pop out' of the search display in a seemingly automatic, purely bottom-up driven fashion. However, research in my laboratory conducted over the last decade has shown that pop-out target detection is subject to dimension-specific processing, or competitive 'weighting', limitations, such that if our visual system is tuned to detecting, say, color-defined targets, its capacity for detecting motion-defined targets is reduced. Also, while these dynamic weighting processes are largely bottom-up driven, they may be modulated by dimensional top-down expectancies. Dimensional weighting in the brain is implemented within a fronto-posterior network of brain areas, with (left) fronto-polar mechanisms initiating the (re-) adjustment of dimensional weight settings, which, via temporal and parietal areas,
modulates the processing efficiency in dimension-specific visual processing areas such as V4 (for color) and hMT+ (for motion). In the lecture, I will present the psychophysical and neuro-scientific evidence for this account and discuss its implications for theories of visual selective attention as well as applications, e.g., in human-robot interaction.
Brief Bio: Hermann Müller is currently Professor of Experimental/Cognitive Psychology at the Universities of Munich, Germany, and London (Birkbeck College). He studied Psychology at the Universities of Wuerzburg, Germany (MSc), and Durham (PhD). After a number of post-doc years (with Professors P.M.A. Rabbitt and G.W. Humphreys), he became a Lecturer/Senior Lecturer/Reader at Birkbeck College, followed by a Professorship at the University of Leipzig (1997-2000). In Munich, he heads a large research group dedicated to the study of selective attention in vision, memory and action. Having worked on all aspects of visual selection (space-based, feature-/dimension-based object-based), over the last decade or so he has become particularly interested in visual search, an activity we engage in every day.
3 February, 2010
Dr. François Pachet, Sony CSL-Paris, 6, rue Aamyot, 75005, Paris, France
Modelling Jazz Virtuosity
Virtuosity can be observed in many domains of human performance, from Jazz improvisation to fast mental computation. This fascinating dimension of human behavior is, however, poorly studied outside of the field of art history.In this talk I focus on the particular problem of generating virtuoso Bebop melodies. The problem of modeling Jazz improvisation has received a lot of attention recently, thanks to progresses in machine learning, statistical modeling, and to the increase in computation power of machines. The Continuator (Pachet, 2003) was the first real time interactive systems to allow users to create musical dialogs using style learning techniques. The Continuator is based on a modeling of musical sequences using Markov chains, a technique that was shown to be well adapted to capture stylistic musical patterns, notably in the pitch domain. The Continuator had great success in free-form improvisational settings, in which the users explore freely musical language created on-the-fly, without additional musical constraints, and was used with Jazz musicians as well as with children (Addessi & Pachet, 2005). However, the Continuator, like most systems using Markovian approaches, is difficult, if not impossible to control. This limitation is intrinsic to the greedy, left-to-right nature of Markovian music generation algorithms. Consequently, it was so far difficult to use these systems in highly constrained musical contexts such as Bebop.
I propose here a computational model of virtuosity based on a novel, combinatorial view of Markov sequence generation. This model solves the "control" problem inherent to Markov chain geeneration, and also provides a very fine degree of control to the user. I illustrate this work with a controlable Bebop improvisation generator. Bebop was chosen as it is a particularly "constrained" style, notably harmonically. I will show how this technique can generate improvisations that satisfy three types of constraints:
(1) harmonic constraints derived from the rules of Bebop,
(2) "Side-slips" as a way to extend the boundaries of Markovian generation by producing locally dissonant but semantically equivalent musical material that smoothly comes back to the authorized tonalities, and 3) non-Markovian constraints deduced from the user’s gestures. I will try to convince the audience that these generated phrases 1) are of the same nature than what real virtuoso are able to generate and that
(3) the ability to control these phrases is an highly enjoyable process.
Brief Bio: Francois Pachet got a Ph.D in Artificial Intelligence from the University of Paris 6. He is now senior researcher at Sony Computer Science Laboratories in Paris, where he conducts research on new forms of musical experiences.
10 February, 2010
Professor Nick Donnelly, University of Southampton
What Mrs Thatcher taught me about face processing.
The Thatcher illusion is thought to demonstrate the processing of perceptual configural features in faces. The basis for this inference is the immediacy of the phenomenological experience of grotesqueness that emerges when the eyes and mouths are inverted in otherwise upright faces. I will report on three experiments that explore how Thatcher faces are discriminated from typical faces. The results of these experiments are inconsistent with perceptual configural processing of Thatcher (and typical) faces. I will go on to argue none of the commonly reported
behavioural tests of configural face processing actually provide supporting evidence for the perceptual processing of configurations. I will finish by considering the implications for studies of both the development of face processing and face processing in atypical populations.
Brief Bio: Nick Donnelly is currently head of the School of Psychology at the University of Southampton. He studied for his PhD at the University of Wales, Swansea, graduating in 1989. Since then he has worked at Birkbeck College, the University of Birmingham, the University of Kent at Canterbury before moving to Southampton in 1999. His research focuses on issues of configurality in visual processing and visual search.
24 February, 2010
Dr. Nadia Berthouze, UCLIC, University College London, UK
Body Movement as a Modality for Affective Human-Computer Interaction
In recent years, we are assisting to the emergence of technology that involves and requires its user to be engaged through their body. This has opened the possibility to better exploit and understand this modality to capture, respond to and regulate the affective experience of its user. We report on our studies aimed at using such modality to induce and recognize affective states in users interacting with technology. In the first part, I will show that technology can capture the quality of experience in its user. I will present a system that dynamically learns to recognize the affective state of people from their posture. While successful experiments have been carried out with acted postures, we are currently testing the system’s ability to detect the more subtle affective states of a computer game player. In the second part of the talk, I will present a model of body movement as a modulator of the experience of the player. Through experiment results, I will show that game controllers affording natural body movements can change the quality of engagement of the player and can induce a more emotional experience.Brief Bio: Since 2006 Nadia Berthouze is a lecturer in the UCL Interaction Centre (UCLIC) at the University of London, a Centre in Human-Computer Interaction. After her PhD (1995) in Computer Science and Bio-medicine at the University of Milan (Italy), she spent 5 years first as a postdoctoral fellow and then as a COE fellow at the Electrotechnical Laboratory (Tsukuba in Japan) where she investigated HCI aspects in the area of Multimedia information interpretation with a focus on the interpretation of affective content. In 2000, she was appointed as lecturer at the Computer Software Department of the University of Aizu in Japan where she extended her interest in emotion expression to the study of non-verbal affective communication. The premise of her research is that affect, emotion, and subjective experience should be factored into the design of interactive technology. At the centre of her research is the creation of interactive systems that exploit body movement as a medium to induce, recognize and measure the quality of experience of humans. She is investigating the various factors involved in the way body movement is used to express and experience emotions, including cross-cultural differences and task context. She was awarded a 2-year International Marie Curie Reintegration Grant (AffectME) to investigate these issues in the clinical domain and in the gaming industry. In the area of clinical domain, she is investigating how to design technology that supports self-directed rehabilitation in chronic muscle-skeleton pain. In the area of computer games, she is investigating how an increase in task-related body movement imposed, or allowed, by the game controller affects the player’s game experience.
3 March, 2010
Dr Michael Proulx, Queen Mary, University of London
Top-down and bottom-up processes in visual search
How does one visually search for a target? Understanding the relative role of top-down and bottom-up guidance is crucial for models of visual search. Here an attentional-capture method was used to address the role of top-down and bottom-up processes in visual search for features and conjunctions. I will discuss two features, brightness and size, as examples of what can capture attention due to the use of bottom-up guidance. I will then turn to why and how attention is captured by these features as a function of the perceptual load or difficulty of the task predicting the reliance on bottom-up mechanisms to guide attention. Next I will present evidence that bottom-up mechanisms are even used in visual search for a conjunction of features. Finally I will also describe recent work considering more complex 'features' that may guide bottom-up mechanisms in visual search, such as size created by the Müller-Lyer illusion.Brief Bio: Michael Proulx first studied psychology at Arizona State University, where he conducted research on categorisation with Donald Homa. He then received an MA and PhD at Johns Hopkins University, where he began research on visual attention and perception under Howard Egeth and Steve Yantis. In Germany he then worked as a postdoc in Düsseldorf and expanded his research in a crossmodal direction, studying sensory substitution devices for blind persons in particular. He is now Lecturer in Cognitive Psychology at Queen Mary University of London.
10 March, 2010
Professor Bob Rafal, MD,
Professor of Clinical Neuroscience and Neuropsychology, School of Psychology, University of Bangor;
Seeing is a verb: Neurological observations on visual awareness
While 'seeing' feels like the passive, immediate and complete appreception of the world before our gaze, we are actually aware, at any one instant, of only a fraction of the information in a scene. Demonstrations of change blindness show that what we experience as a complete and coherent world actually contains just a few objects in the scene and the gist of what could be made accessible to us by exploration. Although we have only a very narrow foveal region of high acuity and sharp colour contrast, the scene is not experienced as being fuzzy or de-saturated in colour around the edges. O'Regan and Noe (2002) propose that visual awareness arises as we gain knowledge of the sensory consequences of action and a mastery of 'sensori-motor contingencies'. Consciousness is not a state we are in - it's something we do as we explore the world. It's what happens when we do something and the world pushes back on our sensory apparatus. It seems self-evident that the perception of objects within a spatial reference frame (and ultimately the abstract conceptualisation of space and magnitude) must arise from the cumulative associations of contingent interactions between motor and sensory cortex: from where else could a developing organism derive them?When an action is made, an 'efference copy' of the motor signal is sent by recurrent collaterals from each level of the motor system to the preceding level from which it received the signal (Sommer & Wurtz, 2008). Here we are considering a very special kind of efference copy - the corollary discharge - recorded in sensory cortex that will be stimulated as a result of movement. One visual area receiving corollary discharge, the intraparietal cortex, has neurons that remap their receptive fields before and after eye movements, and that exhibit particularly interesting properties. First, neurons there remap the entire visual field, regardless of the direction of eye movement or the stimulus location. Secondly, they only remap stimuli that are behaviourally salient, i.e. that demand a high priority for action. I'll summarise observations in neurological patients, and using transcranial magnetic stimulation, that inactivation of this area disrupts saccadic updating of the salience map, and observations in patients with Balint's syndrome that the failure to update the salience map causes the visual scene to disappear. Finally, I'll discuss research that will test the hypothesis that 'seeing' arises from parietal corollary discharge - a prediction of the sensory consequences of action, and the testing of those predictions using the world as an 'external memory'.
Brief Bio: Bob Rafal did his first degree in Biology at the University of Delaware and received his MD from Jefferson Medical College in Philadelphia. He trained in neurology at the University of Oregon, and has taught at Brown University and the University of California, Davis. He has taught at Bangor University since 1999, and is Professor of Clinical Neuroscience and Neuropsychology in the School of Psychology and Consultant Neurologist for the North Wales Brain Injury Service.