2020-01-14
244
Jim Howe
Revised June 2007.
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. They set about the task of developing a programming technology that would enable the use of digital computers as an experimental tool. Over the first four decades of AI's life, a considerable amount of time and effort was given over to the design and development of new special purpose list programming languages, tools and techniques. While the symbolic programming approach dominated at the outset, other approaches such as non-symbolic neural nets and genetic algorithms have 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. Here 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, in the early decades of our work in Edinburgh we subscribed heavily to the view that to make progress towards our goal we would need to understand the nature of the processing at both levels and the relationships between them. For example, some of our work focused 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, such as object recognition, whereas other work dealt primarily with tasks at the sub-symbolic level, including automated assembly of objects from parts, mobile robots, and machine vision for navigation.
|
|
|
Much of AI's accumulating know-how 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 had value beyond its contribution to the achievement of AI's scientific goal. It could 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 integrated AI know-how, methods and techniques with know-how, methods and techniques from other disciplines such as Computer Science and Engineering. This led to the construction of practical applications that replicated expert level decision making or human problem solving, making it more readily available to technical and professional staff in organisations. Today, AI/KBS technology has migrated into a plethora of products of industry and commerce, mostly unbeknown to the users.
