Connectionism, Systematicity, and the Frame Problem

Abstract

This paper investigates connectionism's potential to solve the frame problem. The frame problem arises in the context of modelling the human ability to see the relevant consequences of events in a situation. It has been claimed to be unsolvable for classical cognitive science, but easily manageable for connectionism. We will focus on a representational approach to the frame problem which advocates the use of intrinsic representations. We argue that although connectionism's distributed representations may look promising from this perspective, doubts can be raised about the potential of distributed representations to allow large amounts of complexly structured information to be adequately encoded and processed. It is questionable whether connectionist models that are claimed to effectively represent structured information can be scaled up to a realistic extent. We conclude that the frame problem provides a difficulty to connectionism that is no less serious than the obstacle it constitutes for classical cognitive science. 'It appeared that Newell and Simon were well on their way to fulfilling the prediction they had made in 1958 that 'in a visible future. . . the range of problems (computers) can handle will be coextensive with the range to which the human mind has been applied.' (. . .) Simon's claims fell into place as just another example of the phenomenon which Y. BarHillel had called the fallacy of the successful first step. ' In a talk I gave at RAND, I compared AI to alchemy to make the point. Like the alchemists trying to turn lead into gold, I said, AI had fancy equipment, a few flashy demos, and desperately eager patrons, but they simply had not discovered the right approach to the problem' Hubert Dreyfus describing his evaluation (originally published in 1965) of Newell and Simon's early work in classical AI (Dreyfus and Dreyfits, 1986, pp.6-7).