Information for:

Artificial Intelligence at Edinburgh University : a Perspective

November 1994 Jim Howe

The Nature of Artificial Intelligence

Artificial Intelligence (AI) is an experimental science whose goal is to understand the nature of intelligent thought and action. This goal is shared with a number of longer established subjects such as Philosophy, Psychology and Neuroscience. The essential difference is that AI scientists are committed to computational modelling as a methodology for explicating the interpretative processes which underlie intelligent behaviour, that relate sensing of the environment to action in it. Early workers in the field saw the digital computer as the best device available to support the many cycles of hypothesizing, modelling, simulating and testing involved in research into these interpretative processes, and set about the task of developing a programming technology that would enable the use of digital computers as an experimental tool. A considerable amount of time and effort over the last 35 years or so has been given over to the design and development of new programming languages, tools and techniques. While the symbolic programming approach has dominated, other approaches such as non-symbolic neural nets and genetic algorithms have also featured strongly, reflecting the fact that computing is merely a means to an end, an experimental tool, albeit a vital one.

The popular view of intelligence is that it is associated with high level problem solving, i.e. people who can play chess, solve mathematical problems, make complex financial decisions, and so on, are regarded as intelligent. What we know now is that intelligence is like an iceberg - a small amount of processing activity relates to high level problem solving, that is the part that we can reason about and introspect, but much of it is devoted to our interaction with the physical environment where we are dealing with information from a range of senses, visual, auditory and tactile, and coupling sensing to action, including the use of language, in an appropriate reactive fashion which is not accessible to reasoning and introspection. Using the terms symbolic and sub-symbolic to distinguish these different processing regimes, we take the view that to make progress towards our goal we will need to understand the nature of the processing at both levels and the relationships between them. This maps to our research in that some of our work focuses primarily on symbolic level tasks, in particular, our work on automated reasoning, expert systems and planning and scheduling systems, some aspects of our work on natural language processing, and some aspects of machine vision, e.g. object recognition, whereas other work deals primarily with tasks at the sub-symbolic level, including automated assembly, mobile robots, machine vision for navigation.

Thus far, most of AI's know-how has resulted from work at the symbolic level, modelling mechanisms for performing complex cognitive tasks in restricted domains, for example, diagnosing faults, extracting meaning from utterances, recognising objects in cluttered scenes. But this know-how has value beyond its contribution to the achievement of AI's scientific goal. It can be packaged and made available for use in the work place. This became apparent in the late 1970s and led to an upsurge of interest in applied AI. In the UK, the term Knowledge Based Systems (KBS) was coined for work which involves the integration of AI know-how, methods and techniques with know-how, methods and techniques from other disciplines such as Computer Science and Engineering to construct systems that replicate expert level decision making or human problem solving to make it generally available to technical and professional staff in organisations. Now, AI/KBS technology is migrating into industry and commerce and a wide variety of university departments.

History of AI at Edinburgh

The Department of Artificial Intelligence can trace its origins to a small research group established in a flat at 4 Hope Park Square in 1963 by Donald Michie, then Reader in Surgical Science. During the Second World War, through his membership of Max Newman's code-breaking group at Bletchley Park, Michie had been introduced to computing and had come to believe in the possibility of building machines that could think and learn. By the early 1960s, the time appeared to be ripe to embark on this endeavour. Looking back, there are three discernible periods in the Department's development, each of roughly ten years' duration. The first covers the period from 1963 to the publication of the Lighthill Report by the Science Research Council in l973. During this period, Artificial Intelligence was recognised by the University, first by establishing the Experimental Programming Unit in January 1965 with Michie as Director, and then by the creation of the Department of Machine Intelligence and Perception in October 1966. By then Michie had persuaded Richard Gregory and Christopher Longuet-Higgins, then at Cambridge University and planning to set up a brain research institute, to join forces with him at Edinburgh. Michie's prime interest lay in the elucidation of design principles for the construction of intelligent robots, whereas Gregory and Longuet-Higgins recognized that computational modelling of cognitive processes by machine might offer new theoretical insights into their nature. Indeed, Longuet-Higgins named his research group the Theoretical Section and Gregory called his the Bionics Research Laboratory. During this period there were remarkable achievements in a number of sub-areas of the discipline, including the development of new computational tools and techniques and their application to problems in such areas as assembly robotics and natural language. The POP-2 symbolic programming language which has supported much subsequent UK research and teaching in AI was designed and developed by Robin Popplestone and Rod Burstall. It ran on a multi-access interactive computing system, the second of its kind to be opened in the UK. By 1973, the research in robotics had produced the FREDDY II robot which was capable of assembling objects automatically from a heap of parts. Unfortunately, from the outset of their collaboration these scientific achievements were marred by significant intellectual disagreements about the nature and aims of research in AI and growing disharmony between the founding members of the Department. When Gregory resigned in 1970, the University's reaction was to transform the Department into the School of Artificial Intelligence which was to be run by a Steering Committee. Its three research groups (Jim Howe had taken over responsibility for leading Gregory's group) were given departmental status. At that time, the Metamathematics Unit, which had been set up by Bernard Meltzer to pursue research in automated reasoning, joined the School. By this time, the high level of discord had become known to the School's main sponsors, the Science Research Council. Its reaction was to invite Sir James Lighthill to review the field. Although his report which was published early in 1973 supported AI research related to automation and to computer simulation of neurophysiological and psychological processes, it was highly critical of basic research in foundational areas such as robotics and language processing. Lighthill's report provoked a loss of confidence in AI by the academic community in the UK which persisted for a decade - the so-called "AI Winter".

Since the new School structure had failed to reduce tensions between senior staff, the second period began with an internal review of AI by a University Court Committee. Reporting in 1974, it recommended significant changes. The School structure was scrapped in favour of a single department, now named the Department of Artificial Intelligence, headed initially by Meltzer who retired in 1977 and was replaced by Howe; a separate unit, the Machine Intelligence Research Unit, was set up to accommodate Michie's work. Over this period, the Department's research was dominated by work on automated reasoning, cognitive modelling, children's learning and computation theory (until 1979 when Rod Burstall and Gordon Plotkin left to join the Theory Group in Computer Science). Some achievements included the design and development of the Edinburgh Prolog programming language by David Warren which strongly influenced the Japanese Government's Fifth Generation Computing Project in the 1980s, Alan Bundy's demonstrations of the utility of meta-level reasoning to control the search for solutions to maths problems, and Jim Howe's successful development of computer based learning environments for a range of primary and secondary school subjects, working with both normal and handicapped children.

Unlike its predecessors, the new Department had also committed itself to contribute to undergraduate teaching. Its first course, AI2, a computational modelling course, was launched in 1974/75. This was followed by an introductory course, AI1, in 1978/79. By 1982, it was able to launch its first joint degree, Linguistics with Artificial Intelligence. There were no blueprints for these courses: in each case, the syllabuses had to be carved out of the body of research.

The third period begins with the launch of the Alvey Programme in advanced information technology in 1983. Owing to the increasing number of successful applications of AI technology to practical tasks, in particular expert systems, the negative impact of the Lighthill Report had dissipated. Now, AI was seen as a key information technology, to be fostered through collaborative projects between UK companies and UK universities. The effects on the Department were significant. By taking full advantage of various funding initiatives provoked by the Alvey programme, its academic staff complement increased rapidly from 4 to 15. The accompanying growth in research activity was focused in four areas, Intelligent Robotics, Knowledge Based Systems, Mathematical Reasoning and Natural Language Processing. During the period, the Intelligent Robotics Group undertook collaborative projects in automated assembly, unmanned vehicles and machine vision. It proposed an innovative hybrid architecture for the hierarchical control of reactive robotic devices, and applied it successfully to industrial assembly tasks using a low cost manipulator. In vision, work focused on 3-D geometric object representation, including methods for extracting such information from range data. Achievements included a working range sensor and range data segmentation package. Research in Knowledge Based Systems included design support systems, intelligent front ends and learning environment. The Edinburgh Designer System, a design support environment for mechanical engineers started under Alvey funding, was successfully generalised to small molecule drug design. The Mathematical Reasoning Group prosecuted its research into the design of powerful inference techniques, in particular the development of proof plans for describing and guiding inductive proofs, with applications to problems of program verification, synthesis and transformation, as well as in areas outside Mathematics such as computer configuration and playing bridge. Besides contributing to the Alvey-sponsored speech recognition project in the Centre for Speech Technology Research and to the research work of the Human Communication Research Centre, research in Natural Language Processing spanned projects in the sub-areas of natural language interpretation and generation. Collaborative projects included the implementation of an English language front end to an intelligent planning system, an investigation of the use of language generation techniques in hypertext-based documentation systems to produce output tailored to the user's skills and working context, and exploration of semi-automated editorial assistance such as massaging a text into house style. During this period, the UGC/UFC started the process of assessing research quality. In 1989, and again in 1992, the Department shared a "5" rating with the other departments making up the University's Computing Science unit of assessment.

The Department's postgraduate teaching also expanded rapidly. A masters degree in Knowledge Based Systems which offers specialisms in Foundations of AI, Expert Systems, Intelligent Robotics and Natural Language Processing, was established in 1983, and is now the largest of the Faculty's taught postgraduate courses with 40-50 graduates annually. Many of the Department's complement of about 60 Ph.D. students are drawn from its ranks. At undergraduate level, the most significant development was the launch, in 1987/88, of the joint degree in Artificial Intelligence and Computer Science, with support from the UFC's Engineering and Technology initiative. Subsequently, the modular structure of the course material has enabled the introduction of joint degrees in AI and Mathematics and AI and Psychology, with the target of doubling the number of honours graduates from its courses from 30 to 60 by the end of the century. Recently, the Department also shared an "Excellent" rating awarded by the SHEFC's quality assessment exercise for its teaching provision in the area of Computer Studies.

The start of the fourth decade in the Department's life coincided with the publication, in 1993, of "Realising our Potential", the Government's new strategy for harnessing the strengths of science and engineering to the wealth creation process. For many departments across the UK, the transfer of technology from academia to industry and commerce is unchartered territory. However, from a relatively early stage in the Department's development, there was strong interest in putting AI technology to work outside the laboratory. With financial banking from ICFC, in 1969 Michie and Howe established a small company, called Conversational Software Ltd, to develop and market the POP-2 symbolic programming language. Probably the first AI spin-off company in the world, CSL's POP-2 systems supported work in UK industry and academia for a decade or more, long after CSL ceased to trade. As is so often the case with small companies, the development costs had outstripped market demand. The Department's next exercise in technology transfer was a more modest affair, and was concerned with broadcasting some of the computing tools developed for its work with schoolchildren. In 1981, a small firm, Jessop Microelectronics, was licensed to manufacture and sell the Edinburgh Turtle, a small motorised cart that can be moved around under program control leaving a trace of its path. An excellent tool for introducing programming, spatial and mathematical concepts to young children, over 1000 were sold to UK schools (including 100 supplied to special schools under a DTI initiative). At the same time, with support from Research Machines, Peter Ross and Ken Johnson re-implemented the children's programming language, LOGO, on Research Machines microcomputers. Called RM Logo, for over a decade this has been supplied to educational establishments throughout the UK by Research Machines.

Owing to the explosion of commercial interest in IT in the early 1980s, the Department was bombarded by requests from UK companies for various kinds of technical assistance. For a variety of reasons, not least the Department's modest size at that time, the most effective way of providing this was to set up a separate organisation to support applications oriented R&D. This led to the launch in July 1983 of the Artificial Intelligence Applications Institute. Its mission is to help its clients acquire know-how and skills in the construction and application of knowledge based systems technology, enabling them to support their own product or service developments and so gain a competitive edge. In practice, the Institute has been a technology transfer experiment: there was no blueprint, no model to specify how the transfer of AI technology could best be achieved. So, much time and effort has been given over to conceiving, developing and testing a variety of mechanisms through which knowledge and skills could be imparted to clients. Ten years on, the Institute employs about twenty technical staff, with an annual turnover just short of 1M. Outside the UK, it has major clients in Japan and the US. Its work is focused on three sub-areas of knowledge-based systems, planning and scheduling systems, decision support systems and information systems. While the bulk of its work is applied, the Institute sustains its own internal strategic research programme in areas of potential industrial demand. In planning, the goal is to develop a domain-independent reactive planner capable of accepting descriptions of planning domains and generating realistic plans for subsequent execution. Currently, the Institute is contributing to a large US research initiative which is exploring the application of knowledge based planning methods to military logistic applications. In the decision support area, the Institute was recently awarded its largest single contract, worth about 2.4M over 3 years. In partnership with a number of UK companies, it will be developing a computer tool set to model, analyse and improve aspects of how a business works and how it is organised.

Within the Department, the research objective for the next decade is to consolidate and strengthen existing activities. In Intelligent Robotics, the hybrid systems architecture will be thoroughly evaluated on realistic industrial assembly and real time tasks. Work on unmanned vehicles will include further development of active vision systems, such as the existing kinetic depth system; investigating the utility of connectionist processing of the information provided by such systems, and exploring the usefulness of biological models of animal behavioural selection and learning for mobile robot control. Research in 3-D object recognition systems will be extended to the recognition and location of complex industrial parts and automatic acquisition of CAD models from shown examples. The problem of grasping unknown objects, for tasks such as radioactive waste disposal, will be investigated. In the Knowledge Based Systems area research will continue into the processes underlying complex decision-making tasks such as design, planning and scheduling. Recently begun work on knowledge-based requirements capture for Software Engineering will be an important research activity, as will the realisation of hybrid knowledge-based systems which combine connectionist and symbolic computing technology. Research work in Mathematical Reasoning will continue, with further testing of the inductive theorem proving system and refinement and extension to the proof plans which guide it. Solutions are anticipated for some hard problems in automated programming and mathematics. The software tools will be developed to suit them better to industrial application. In Natural Language Processing, effort will be focused on the extraction and manipulation of information via natural language. This will cover both trawling for information from databases (including active computer systems, such as expert systems or intelligent planners), and massaging it into forms readily understandable by users. 


Home : About 

School of Informatics, Appleton Tower, Crichton Street, Edinburgh, EH8 9LE, Scotland, UK
Tel: +44 131 650 2690, Fax: +44 131 651 1426, E-mail: hod@inf.ed.ac.uk
Please contact our webadmin with any comments or corrections.
Unless explicitly stated otherwise, all material is copyright © The University of Edinburgh