When Almost Is Not Even Close: Remarks on the Approximability of HDTP

Abstract

This is the accepted version of the paper.This version of the publication may differ from the final published version. Abstract. A growing number of researchers in Cognitive Science advocate the thesis that human cognitive capacities are constrained by computational tractability. If right, this thesis also can be expected to have far-reaching consequences for work in Artificial General Intelligence: Models and systems considered as basis for the development of general cognitive architectures with human-like performance would also have to comply with tractability constraints, making in-depth complexity theoretic analysis a necessary and important part of the standard research and development cycle already from a rather early stage. In this paper we present an application case study for such an analysis based on results from a parametrized complexity and approximation theoretic analysis of the Heuristic Driven Theory Projection (HDTP) analogy-making framework. Permanent

On Cognitive Aspects of Human-Level Artificial Intelligence

“…Complementing these efforts and completing the picture, and Besold and Robere (2013) provide an investigation into the computational complexity and approximability of the approach. 3 Viewing HDTP from the perspective of theoretical computer science as a tool for the computation of generalizations between domain theories by means of a special form of anti-unification (i.e., putting it under the umbrella of rewriting research), properties, such as computational complexity and approximability, will not necessarily play a central role in the evaluation of the overall approach and system.…”

“…The advantage of this form of high-level characterization of limiting factors for models and theories lies in its generality and independence from particular implementation details, by this overcoming many of the aforementioned divisions and barriers within HLAI and cognitive systems research and allowing for the introduction of general principles and requirements. The work by and Besold and Robere (2013) moreover serves as a first worked out example of how a corresponding system analysis can look like and what kinds of results can, among others, be expected.…”

“…We therefore in our future work hope to develop the overall framework further, also showing the usefulness and applicability of the proposed methods in different worked examples from several relevant fields: The range of eligible application scenarios spans from models of epistemic reasoning and interaction, over cognitive systems in general problem-solving scenarios, down to models for particular cognitive capacities as, for example, analogy-making (see, e.g., Sect. 4 and additionally Besold and Robere (2013b) for a proof of concept).…”