The Research Themes are involved in various projects such as facilitating workshops for faculty and research students around a particular topic in order to develop a research programme, bringing together faculty from different departments and schools to respond to exernal funding calls, hosting conferences and seminars, and hosting academic journals.
The following list highlights some of the externally funded Mind and Brain projects currently being undertaken by Sussex Faculty.
- A Unified Model of Compositional and Distributional Semantics: Theory and Applications
Dr David J Weir, Engineering and Physical Sciences Research Council
There have been two main approaches to modelling the meaning of language in NLP. The first, the so-called compositional approach, is based on classical ideas from Philosophy and Mathematical Logic. Using a well-known principle from the 19th century logician Frege - that the meaning of a phrase can be determined from the meanings of its parts and how those parts are combined - logicians have developed formal accounts of how the meaning of a sentence can be determined from the relations of words in a sentence. This idea culminated famously in Linguistics in the work of Richard Montague in the 1970s. The compositional approach addresses a fundamental problem in Linguistics - how it is that humans are able to generate an unlimited number of sentences using a limited vocabulary. We would like computers to have a similar capacity also. The second, more recent, approach to modelling meaning in NLP focuses on the meanings of the words themselves. This is the so-called distributional approach to modelling word meanings and is based on the ideas of the "structural" linguists such as Firth from the 1950s. This idea is also sometimes related to Wittenstein's philosophy of "meaning as use". The idea is that the meanings of words can be determined by considering the contexts in which words appear in text. For example, if we take a large amount of text and see which words appear close to the word "dog", and do a similar thing for the word "cat", we will see that the contexts of dog and cat tend to share many words in common (such as walk, run, furry, pet, and so on). Whereas if we see which words appear in the context of the word "television", for example, we will find less overlap with the contexts for "dog". Mathematically we represent the contexts in a vector space, so that word meanings occupy positions in a geometrical space. We would expect to find that "dog" and "cat" are much closer in the space than "dog" and "television", indicating that "dog" and "cat" are closer in meaning than "dog" and "television". The two approaches to meaning can be roughly characterized as follows: the compositional approach is concerned with how meanings combine, but has little to say about the individual meanings of words; the distributional approach is concerned with word meanings, but has little to say about how those meanings combine. This project exploits the strengths of the two approaches, by developing a unified model of distributional and compositional semantics. The project has a central theoretical component, drawing on models of semantics from Theoretical Computer Science and Mathematical Logic. This central component which will inform, be driven by, and evaluated on tasks and applications in NLP and Information Retrieval, and also data drawn from empirical studies in Cognitive Science (the computational study of the mind). Hence we aim to make the following contributions: (i) advance the theoretical study of meaning in Linguistics, Computer Science and Artificial Intelligence; (ii) develop new meaning-sensitive approaches to NLP applications which can be robustly applied to naturally occurring text.
This multisite project is funded by the EPSRC, with related grants funding the project at the Universities of Cambridge, Edinburgh, Oxford and York.
- Green Brain: Computational Modelling of the Honeybee Brain
Dr Thomas Nowotny, Dr James Marshall, Dr Eleni Vasilaki, Professor Kevin Gurney, Engineering and Physical Sciences Research Council
The development of an ‘artificial brain’ is one of the greatest challenges in artificial intelligence, and its success will have innumerable benefits in many and diverse fields, from robotics to cognitive psychology. Most research effort is spent on modelling vertebrate brains. Yet smaller brains can display comparable cognitive sophistication, while being more experimentally accessible and amenable to modelling. The ‘Green Brain’ project will combine computational neuroscience modelling, learning and decision theory, modern parallel computing methods and robotics with data from state-of-the-art neurobiological experiments on cognition in the honeybee Apis mellifera, to build and deploy a modular model of the honeybee brain describing detection, classification and learning in the olfactory and optic pathways as well as multi-sensory integration across these sensory modalities. Unlike other brain models which use expensive traditional supercomputing resources, the ‘Green Brain’ will be implemented on massively parallel, but affordable GPU technology. The ‘Green Brain’ will be deployed for the real-time control of a flying robot able to sense and act autonomously; this robot testbed will be used to demonstrate the development of new biomimetic control algorithms for artificial intelligence and robotics applications. Further, by modelling complete sensorimotor loops endowed with behaviour, we will be able to begin examining the nature of embodied cognition in biological brains rather than abstract agents.
- The role of vicarious learning in preventing and treating children's fears
The role of vicarious learning in preventing and treating children's fears
Joint project with Kingston University, Dr Chris Askew (PI) Professor Andy P Field, Economic and Social Research Council
Anxiety is the most common of all childhood psychological disorders. It frequently hinders children's social and academic functioning and can therefore lead to problems in later life. Children can learn to become frightened of certain objects if they witness someone else acting frightened of them. This process is called "vicarious learning". To develop the most effective prevention and treatment programmes for fears acquired in this way it is necessary to understand when and how vicarious learning occurs by investigating the mechanism underpinning it. This is the first aim of the research programme. As with other anxiety disorders, clinical fears and phobias are associated with specific patterns of responses toward the feared object, including: avoiding it, saying negative things about it, and increased heart rate. Researchers have found that vicarious learning can cause changes in the first two types of response, but to better understand phobias it's important to show changes in heart rate too. Researchers have also shown that anxiety is typically maintained by biased ways of looking and thinking; in particular, anxious individuals tend to notice and pay more attention to their feared object (attentional bias). However, although it is thought likely that these biases begin in childhood, it is not yet known how or why they develop. Hence, we still know relatively little about this very important component of anxiety, therefore limiting our ability to prevent anxiety from developing in children. One probable scenario is that these biases begin following learning experiences with the feared object. The second aim of the research is therefore to see whether vicarious learning can affect heart rate responses to an object and how much attention is paid to it. A third aim is to evaluate ways of preventing and treating the negative outcomes of vicarious learning. Recent studies show that anxiety is reduced when people are trained to divert their attention away from the object of their fear (attentional training). We will examine whether vicariously learnt fear responses towards animals can be reversed using this type of training. Finally, we will investigate whether vicarious learning can also be used to protect against acquiring fears and eliminate existing fears. This work is of interest to psychologists, parents and teachers. It will offer guidance on how to minimise the risk of transmitting fears to children and suggest ways to intervene quickly when it is recognised that a child may have been involved in a fear-related vicarious learning event. Finally, the research also has wider implications for how animals and other stimuli are presented to children (e.g. on television, in films, or in books).
- Diagnosing vulnerability and ‘dangerousness’: police decision-making in the implementation of Section 136
Professor Gillian A Bendelow, BA Senior Research Fellowship in collaboration with Sussex Partnership NHS Foundation Trust and Sussex Police. British Academy
Public anti-social behaviour is increasingly a major societal concern and police in England and Wales are empowered, under Section 136 of the Mental Health Act 1983, to detain individuals who are thought to be a danger to themselves or to others. Use of this authority is widespread, but attracts controversy, as it requires the police to make judgements about mental health. Detention under section 136 in Brighton and Hove is approximately ten times the national average, but over the last year, less than one third of police detainees in public places were diagnosed as having a 'treatable' mental illness. This research ties together previous theoretical and emprical work by tracking and recording the decision- making processes of both police and mental health professionals involved in ALL s136 case over one year in Sussex, using in-depth interviews and observational techniques. Findings would produce policy implications to address seemingly intractable problems at both local and national levels, as well as contributing to current diagnostic debates around dangerous and vulnerable 'personality disorders'.
- CATEGORIES: The origin and impact of colour categories in language and thought
Dr Anna Franklin, European Union
Humans can discriminate millions of colours, yet language refers to colour using a number of discrete categories (e.g., red, green, blue). These colour categories are also present in thought (e.g., in colour judgements / memory). There has been considerable multidisciplinary research into the origin of colour categories and how colour categories in thought and language relate. However, major theoretical challenges remain. The 5 year 'CATEGORIES' project, led by Franklin, will tackle these crucial challenges. In our previous research, we have established that infants respond categorically to colour. The first aim of the ‘CATEGORIES’ project is to establish the relationship of these pre-linguistic colour categories to the commonality and variation in the world’s colour lexicons. In order to achieve this, we are conducting a series of sub-projects which draw on a diverse range of methods (e.g., infant testing, computational simulations, psychophysics, cross-cultural fieldwork). The second aim is to resolve the debate about the effect of colour terms on colour perception. It has previously been claimed that speakers of different colour lexicons see colour differently, a proposal which relates to Whorf’s hypothesis (1956) that language influences our perception of the world. We are developing a ‘Neuro-Whorfian’ approach, using neuro-physiological methods to make cross-linguistic comparisons of colour processing. Overall, the project aims to provide new questions, approaches, data and theory to the multidisciplinary field of colour category research. More broadly, the project addresses issues that are fundamental to understanding the complexity of the human mind, such as the interaction of language and thought, and how the brain categorises the visual world.
- EXPECT_CONSCIOUS: When do expectations affect conscious perception?
Professor Anil K Seth and Dr. Yair Pinto, European Union
There is accelerating interest in the role of expectations (predictions) in cognition and behaviour, under the general rubric of ‘predictive coding’. Abundant evidence now demonstrates influences of expectations, at both behavioural and neural levels, in sensorimotor performance and perception. A version of predictive coding has even been suggested as a fundamental principle underlying all aspects of brain operation. However, the influence of expectations specifically on conscious perception remain largely unknown. The present project, a Marie Curie fellowship awarded to Prof Seth and Dr. Yair Pinto, who joins from the University of Amsterdam, addresses this issue head on. The team will ask under what conditions can stimuli induce expectations that influence subsequent conscious perception. Specifically, they will use a combination of psychophysics and neuroimaging to examine (i) the role of ‘top-down’ versus ‘bottom-up’ expectations, and (ii) whether attention and/or subjective visibility is necessary such that expectation-inducing stimuli affect subsequent perceptions. Behavioural experiments will expose the functional architecture of interactions between expectations and conscious perception, and neuroimaging studies will reveal aspects of the underlying neural mechanisms. Combining these approaches will allow us to situate the work within – and help develop – mechanistic contexts provided by predictive coding within cortical hierarchies.
EXPECT_CONSCIOUS fits within in a larger research programme, pursued within the Sackler Centre for Consciousness Science, aimed at understanding the roles of prediction and expectation in conscious experience. Visit the Sackler Centre website for more information.