Mark Bishop

Goldsmiths, University of London, Computing, Faculty Member

Followers

212

Following

Co-authors

Public Views

Cognitive Computing is an interdisciplinary subject involving biologically-inspired computational accounts of all aspects of natural and artificial cognitive systems and bridges the gap between life sciences, social sciences, engineering, physical and mathematical sciences, and humanities. Areas of particular interest include: artificial intelligence, neural networks, cognitive neuromorphic engineering and other hardware implementations, the philosophy of cognitive computation, cognitive robotics, autonomous cognitive systems, neuroscience nanotechnology, self-organizing, swarm and immune systems, complex systems and control theory, and computational cognitive neuroscience, as well as focusing on the development of latest research into practical Cognitive Computing applications.

Under the aegis of Cognitive Computing my research activity spans Computational Intelligence (theory and applications) and the Philosophy of Computational Intelligence.
Phone: +44 207 078 5048
Address: Room 2, 29 St. James,
New Cross,
LONDON.
SE14 6NW.

less

InterestsView All (26)

Uploads

Papers by Mark Bishop

Heightened pain sensitivity in individuals with signs and symptoms of carpal tunnel syndrome and the relationship to clinical outcomes following a manual therapy intervention

Manual Therapy, 2011

Neurophysiological responses related to lessening of pain sensitivity are a suggested mechanism o... more Neurophysiological responses related to lessening of pain sensitivity are a suggested mechanism of manual therapy. Prior studies have observed generalized lower pain thresholds associated with carpal tunnel syndrome (CTS) in comparison to healthy controls. The present study sought to determine whether similar findings were present in suprathreshold measures and measures specific to central integration of pain (temporal summation and after sensations). Additionally, we wished to determine whether measures of pain sensitivity were related to clinical outcomes in participants with signs and symptoms of CTS receiving a manual therapy intervention. Individuals with signs and symptoms of CTS reported greater pain sensitivity to suprathreshold measures of mechanical pain, temporal summation, and after sensation in comparison to healthy controls. Immediate lessening of mechanical pain sensitivity and after sensations in response to a manual therapy intervention and 3-week attenuation of temporal summation following a 3-week course of manual therapy were associated with 3-week changes in clinical pain intensity in participants with signs and symptoms of CTS. These findings suggest heightened pain sensitivity across several parameters may be associated with CTS. Furthermore, changes in mechanical pain, after sensation, and temporal summation may be related to improvements in clinical outcomes.

Download

Review - Artificial Intelligence: a modern myth?

Book review ARTIFICIAL INTELLIGENCE: AModern Myth John Kelly £34.95, 253 pp. ISBN 0-13-338559-0 (hb

Stochastic diffusion search

Scholarpedia, 2007

Stochastic Diffusion Search is a well characterised robust swarm intelligence global metaheuristi... more

History and Philosophy of Neural Networks

This chapter conceives the history of neural networks emerging from two millennia of attempts to ... more This chapter conceives the history of neural networks emerging from two millennia of attempts to rationalise and formalise the operation of mind. It begins with a brief review of early classical conceptions of the soul, seating the mind in the heart; then discusses the subsequent Cartesian split of mind and body, before moving to analyse in more depth the twentieth century hegemony identifying mind with brain; the identity that gave birth to the formal abstractions of brain and intelligence we know as 'neural networks'. The chapter concludes by analysing this identity-of intelligence and mind with mere abstractions of neural behaviour-by reviewing various philosophical critiques of formal connectionist explanations of 'human understanding', 'mathematical insight' and 'consciousness'; critiques which, if correct, in an echo of Aristotelian insight, suggest that cognition may be more profitably understood not just as a result of [mere abstractions of] neural firings, but as a consequence of real, embodied neural behaviour, emerging in a brain, seated in a body, embedded in a culture and rooted in our world; the so called 4Es approach to cognitive science: the Embodied, Embedded, Enactive, and Ecological conceptions of mind. Contents 1. Introduction: the body and the brain 2. First steps towards modelling the brain 3. Learning: the optimisation of network structure 4. The fall and rise of connectionism 5. Hopfield networks 6. The 'adaptive resonance theory' classifier 7. The Kohonen 'feature-map' 8. The multi-layer perceptron 9. Radial basis function networks 10. Recent developments in neural networks 11. "What artificial neural networks cannot do .." 12. Conclusions and perspectives Glossary Nomenclature References Biographical sketches

Download

AISB50: Artificial Intelligence at a New Branch Point

Download

Cooperation of Nature and Physiologically Inspired Mechanisms in Visualisation

Biologically-Inspired Computing for the Arts

A novel approach of integrating two swarm intelligence algorithms is considered, one simulating t... more A novel approach of integrating two swarm intelligence algorithms is considered, one simulating the behaviour of birds flocking (Particle Swarm Optimisation) and the other one (Stochastic Diffusion Search) mimics the recruitment behaviour of one species of ants – Leptothorax acervorum. This hybrid algorithm is assisted by a biological mechanism inspired by the behaviour of blood flow and cells in blood vessels, where the concept of high and low blood pressure is utilised. The performance of the nature-inspired algorithms and the biologically inspired mechanisms in the hybrid algorithm is reflected through a cooperative attempt to make a drawing on the canvas. The scientific value of the marriage between the two swarm intelligence algorithms is currently being investigated thoroughly on many benchmarks, and the results reported suggest a promising prospect (al-Rifaie, Bishop & Blackwell, 2011). It may also be discussed whether or not the artworks generated by nature and biologically ...

Download

The autonomous mobile robot SENARIO: a sensor aided intelligent navigation system for powered wheelchairs

IEEE Robotics & Automation Magazine, 1997

The SENARIO project is develoing a sensor-aided intelligent navigation system that provides high-... more The SENARIO project is develoing a sensor-aided intelligent navigation system that provides high-level navigational aid to users of powered wheelchairs. The authors discuss new and improved technologies developed within SENARIO concerning task/path planning, sensing and positioning for indoor mobile robots as well as user interface issues. The autonomous mobile robot SENARIO, supports semi- or fully autonomous navigation. In semi-autonomous mode

The mining game: a brief introduction to the stochastic diffusion search metaheuristic

Q: The magazine of AISB, 2010

In recent years studies of social agents have suggested several new metaheuristics for use in sea... more In recent years studies of social agents have suggested several new metaheuristics for use in search and optimisation; Stochastic Diffusion Search (SDS)[1] is one such 'Swarm Intelligence'algorithm. SDS is a distributed population based search algorithm utilising interaction between simple agents to locate a global optimum; such 'communicating agents' have recently been suggested as a potential metaphor for some cognitive processes [6]. SDS is most easily applied to discrete search and optimisation problems where the task is ...

Download

Evolutionary training of artificial neural networks for radiotherapy treatment of cancers

1998 IEEE International Conference on Evolutionary Computation Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98TH8360)

Abstract Radiotherapy treatment of cancers involves the need to orient a set of X-ray beams direc... more Abstract Radiotherapy treatment of cancers involves the need to orient a set of X-ray beams directed at a tumour in such a way as to maximise the dosage of X-rays to the tumour itself, while minimising the effect on surrounding healthy tissue and organs. This process is either planned by a radiotherapist, which is a time-consuming process often with suboptimal results, or rarely, where expensive computing equipment is available, by direct optimisation using sophisticated search techniques, which is similarly time-consuming. This paper ...

Classification in e-procurement

International Journal of Applied Pattern Recognition, 2014

Three coupled knowledge transfer partnerships used pattern recognition techniques to produce an e... more Three coupled knowledge transfer partnerships used pattern recognition techniques to produce an e-procurement system which, the National Audit Office reports, could save the National Health Service £500 m per annum. An extension to the system, GreenInsight, allows the environmental impact of procurements to be assessed and savings made. Both systems require suitable products to be discovered and equivalent products recognised, for which classification is a key component. This paper describes the innovative work done for product classification, feature selection and reducing the impact of mislabelled data.

Download

Entity Search/Match in Relational Databases

Proceedings of the 9th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management, 2017

We study an entity search/match problem that requires retrieved tuples match to an input entity q... more We study an entity search/match problem that requires retrieved tuples match to an input entity query. We assume the input queries are of the same type as the tuples in a materialised relational table. Existing keyword search over relational databases focuses on assembling tuples from a variety of relational tables in order to respond to a keyword query. The entity queries in this work differ from the keyword queries in two ways: (i) an entity query roughly refers to an entity that contains a number of attribute values, i.e. a product entity or an address entity; (ii) there might be redundant or incorrect information in the entity queries that could lead to misinterpretations of the queries. In this paper, we propose a transformation that first converts an unstructured entity query into a multi-valued structured query, and two retrieval methods are proposed to generate a set of candidate tuples from the database. The retrieval methods essentially formulate SQL queries against the database given the multi-valued structured query. The results of a comprehensive evaluation of a large-scale database (more than 29 millions tuples) and two real-world datasets showed that our methods have a good trade-off between generating correct candidates and the retrieval time compared to baseline approaches.

Download

Convergence of the Stochastic Diffusion Search

In this paper we present a connectionist searching technique- the Stochastic Diffusion Search (SD... more In this paper we present a connectionist searching technique- the Stochastic Diffusion Search (SDS), capable of rapidly locating a specified pattern in a noisy search space. In operation SDS finds the position of the prespecified pattern or if it does not exist - its best instantiation in the search space. This is achieved via parallel exploration of the whole search space by an ensemble of agents searching in a competitive cooperative manner. We prove mathematically the convergence of stochastic diffusion search. SDS converges to a statistical equilibrium when it locates the best instantiation of the object in the search space. Experiments presented in this paper indicate the high robustness of SDS and show good scalability with problem size. The convergence characteristics of SDS makes it a fully adaptive algorithm and suggests applications in dynamically changing environments.

HeX and the Single Anthill: Playing Games with Aunt Hillary

Fundamental Issues of Artificial Intelligence, 2016

In a reflective and richly entertaining piece from 1979, Doug Hofstadter playfully imagined a con... more In a reflective and richly entertaining piece from 1979, Doug Hofstadter playfully imagined a conversation between 'Achilles' and an anthill (the eponymous 'Aunt Hillary'), in which he famously explored many ideas and themes related to cognition and consciousness. For Hofstadter, the anthill is able to carry on a conversation because the ants that compose it play roughly the same role that neurons play in human languaging; unfortunately, Hofstadter's work is notably short on detail suggesting how this magic might be achieved 1. Conversely in this paper-finally reifying Hofstadter's imagination-we demonstrate how populations of simple ant-like creatures can be organised to solve complex problems; problems that involve the use of forward planning and strategy. Specifically we will demonstrate that populations of such creatures can be configured to play a strategically strongthough tactically weak-game of HeX (a complex strategic game). We subsequently demonstrate how tactical play can be improved by introducing a form of forward planning instantiated via multiple populations of agents; a technique that can be compared to the dynamics of interacting populations of social insects via the concept of meta-population. In this way although, pace Hofstadter, we do not establish that a meta-population of ants could actually hold a conversation with Achilles, we do successfully introduce Aunt Hillary to the complex, seductive charms of HeX.

Download

Extending Naive Bayes Classifier with Hierarchy Feature Level Information for Record Linkage

Lecture Notes in Computer Science, 2015

Probabilistic record linkage has been well investigated in recent years. The Fellegi-Sunter proba... more Probabilistic record linkage has been well investigated in recent years. The Fellegi-Sunter probabilistic record linkage and its enhanced version are commonly used methods, which calculate match and non-match weights for each pair of corresponding fields of record-pairs. Bayesian network classifiers-naive Bayes classifier and TAN have also been successfully used here. Very recently, an extended version of TAN (called ETAN) has been developed and proved superior in classification accuracy to conventional TAN. However, no previous work has applied ETAN in record linkage and investigated the benefits of using a naturally existing hierarchy feature level information. In this work, we extend the naive Bayes classifier with such information. Finally we apply all the methods to four datasets and estimate the F1 scores.

Download

Second‐order cybernetics and enactive perception

Kybernetes, 2005

PurposeTo present an account of cognition integrating second‐order cybernetics (SOC) together wit... more PurposeTo present an account of cognition integrating second‐order cybernetics (SOC) together with enactive perception and dynamic systems theory.Design/methodology/approachThe paper presents a brief critique of classical models of cognition then outlines how integration of SOC, enactive perception and dynamic systems theory can overcome some weaknesses of the classical paradigm.FindingsPresents the critique of evolutionary robotics showing how the issues of teleology and autonomy are left unresolved by this paradigm although their solution fits within the proposed framework.Research limitations/implicationsThe paper highlights the importance of genuine autonomy in the development of artificial cognitive systems. It sets out a framework within which the robotic research of cognitive systems could succeed.Practical implicationsThere are no immediate practical implications but see research implications.Originality/valueIt joins the discussion on the fundamental nature of cognitive sys...

Download

Application of neural networks to computer recipe prediction

Color Research & Application, 1991

Conventional mechanisms for computer colorant formulation commonly employ the Kubelka-Munk theory... more Conventional mechanisms for computer colorant formulation commonly employ the Kubelka-Munk theory to relate reflectance values to colorant concentrations, however there are situations where this approach is not applicable and hence an alternative is desirable. One such method is to utilise Artificial Intelligence techniques to mimic the behaviour of the professional colourist. The purpose of this paper is to describe recent research being carried out at Reading University, sponsored by Courtaulds Research, that utilises collections of cellular automata, known as Neural Networks, in the problem of recipe prediction.

Download

An efficient pruning technique for the global optimisation of multiple gravity assist trajectories

Napsac: High Noise, High Dimensional Model Parameterisation-It’s in the Bag

Abstract. The most effective algorithms for model parameterisation in the presence of high noise,... more Abstract. The most effective algorithms for model parameterisation in the presence of high noise, such as RANSAC, MINPRAN and Least Median Squares, use random sampling of data points to instantiate model hypotheses. However, their performance degrades in higher dimensionality due to the exponentially decreasing probability of sampling a set of inliers. It is suggested that biasing this random selection towards clusters in multi-dimensional space may be preferable. Based on this premise, the NAPSAC (N Adjacent Points Sample ...

Download

Stochastic diffusion optimisation: the application of partial function evaluation and stochastic recruitment in swarm intelligence optimisation

The concept of partial evaluation of ¯tness functions, together with mechanisms manipulating the ... more The concept of partial evaluation of ¯tness functions, together with mechanisms manipulating the resource allocation of population based search methods, are presented in the context of Stochastic Diffusion Search, a novel swarm intelligence metaheuristic that has many similarities with ant and evolutionary algorithms. It is demonstrated that the stochastic process ensuing from these algorithmic concepts has properties that allow the algorithm to optimise noisy fitness functions, to track moving optima, and to redistribute the population after quantitative changes in the fitness function. Empirical results are used to validate theoretical arguments.

The Stochastic Search Network

Neural Networks for Vision, Speech and Natural Language, 1992

ABSTRACT A fundamental difficulty when using neural networks applied to pattern recognition, is t... more ABSTRACT A fundamental difficulty when using neural networks applied to pattern recognition, is the problem of stimulus equivalence - the invariance of symbolic information independent of transformation within a search space. For example, symbolically, the letter A remains an A irrespective of its position, size or orientation within an image field. Adult humans can generally recognise patterns accurately, despite such transformations and distortions. Classically it has been hypothesised that this ability is due to a normalisation process that occurs before the classification process begins. One criticism of the network paradigm has been that there is no clear mechanism by which such normalisation can occur, since at first sight such unstructured systems do not provide any mechanism for attention - the focussing of analysis onto a subset of a larger pattern. This chapter will describe two network solutions to this normalisation process: Hinton Mapping and Stochastic Networks, give an introduction to the analysis of Stochastic Networks and illustrate their use within the British Telecom CONNEX project. The problem of stimulus equivalence is a specific example of a best fit constraint satisfaction search, where the target pattern has been mapped into the search space, using translational, rotational or scale transformations. The goal of a system that can solve such problems is to ascertain the inverse mapping from search space to pattern. For example, in the situation where a letter has been rotated in a visual field, the inverse mapping would be the angle of rotation q, if it had been translated then the inverse mapping would be its cartesian coordinates X,Y, etc. One solution to this problem is a technique described by Hinton, summarised below.

Mark Bishop

Uploads

Papers by Mark Bishop

Log In