Ahmed M. H. Abdel-Fattah

Algorithmic Aspects of Theory Blending

In Cognitive Science, conceptual blending has been proposed as an important cognitive mechanism that facilitates the creation of new concepts and ideas by constrained combination of available knowledge. It thereby provides a possible theoretical foundation for modeling high-level cognitive faculties such as the ability to understand, learn, and create new concepts and theories. This paper describes a logic-based framework which allows a formal treatment of theory blending, discusses algorithmic aspects of blending within the framework, and provides an illustrating worked out example from mathematics.

Creativity, cognitive mechanisms, and logic

Creativity is usually not considered to be a major issue in current AI and AGI research. In this paper, we consider creativity as an important means to distinguish human-level intelligence from other forms of intelligence (be it natural or artificial). We claim that creativity can be reduced in many interesting cases to cognitive mechanisms like analogy-making and concept blending. These mechanisms can best be modeled using (non-classical) logical approaches. The paper argues for the usage of logical approaches for the modeling of manifestations of creativity in order to step further towards the goal of building an artificial general intelligence.

Theory Blending as a Framework for Creativity in Systems for General Intelligence

Being creative is a central property of humans in solving problems, adapting to new states of affairs, applying successful strategies in previously unseen situations, or coming up with new conceptualizations. General intelligent systems should have the potential to realize such forms of creativity to a certain extent. We think that creativity and productivity issues can be best addressed by taking cognitive mechanisms into account, such as analogy making, concept blending, computing generalizations and the like.

Utilizing cognitive mechanisms in the analysis of counterfactual conditionals by AGI systems

We give a crisp overview of the problem of analyzing counterfactual conditionals, along with a proposal of how an artificial system can overcome its challenges, by operationally utilizing computationally-plausible cognitive mechanisms. We argue that analogical mapping, blending of knowledge from conceptual domains, and utilization of simple cognitive processes lead to the creative production of, and the reasoning in, mentally-created domains, which shows that the analysis of counterfactual conditionals can be done in computational models of general intelligence.

Rationality and general intelligence

Humans are without any doubts the prototypical example of agents that can hold rational beliefs and can show rational behavior. If an AGI system is intended to model the full breadth of human-level intelligence, it is reasonable to take the remarkable abilities of humans into account with respect to rational behavior, but also the apparent deficiencies of humans in certain rationality tasks. Based on well-known challenges for human rationality (Wason-Selection task and Tversky & Kahnemans Linda problem) we propose that rational belief of humans is based on cognitive mechanisms like analogy making and coherence maximization of the background theory. The analogy making framework Heuristic-Driven Theory Projection (HDTP) can be used for implementing these cognitive mechanisms.

Employing a Restricted Set of Qualitative Relations in Recognizing Plain Sketches

In this paper, we employ aspects of machine learning, computer vision, and qualitative representations to build a classifier of plain sketches. The paper proposes a hybrid technique for accurately recognizing hand-drawn sketches, by relying on a set of qualitative relations between the strokes that compose such sketches, and by taking advantage of two major perspectives for processing images. Our implementation shows promising results for recognizing sketches that have been hand-drawn by human participants.

Theory blending: extended algorithmic aspects and examples

In Cognitive Science, conceptual blending has been proposed as an important cognitive mechanism that facilitates the creation of new concepts and ideas by constrained combination of available knowledge. It thereby provides a possible theoretical foundation for modeling high-level cognitive faculties such as the ability to understand, learn, and create new concepts and theories. Quite often the development of new mathematical theories and results is based on the combination of previously independent concepts, potentially even originating from distinct subareas of mathematics. Conceptual blending promises to offer a framework for modeling and re-creating this form of mathematical concept invention with computational means. This paper describes a logic-based framework which allows a formal treatment of theory blending (a subform of the general notion of conceptual blending with high relevance for applications in mathematics), discusses an interactive algorithm for blending within the framework, and provides several illustrating worked examples from mathematics.

Model construction in general intelligence

In this conceptual paper we propose a shift of perspective for parts of AI --- from search to model construction. We claim that humans construct a model of a problem situation consisting of only a few, hierarchically structured elements. A model allows to selectively explore possible continuations and solutions, and for the flexible instantiation and adaptation of concepts. We underpin our claims by results from two protocol studies on problem-solving imagery and on the inductive learning of an algorithmic structure. We suggest that a fresh look into the small-scale construction processes humans execute would further ideas in categorization, analogy, concept formation, conceptual blending, and related fields of AI.