2379-8920 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TCDS.2016.2629622, IEEE Transactions on Cognitive and Developmental Systems IEEE TRANSACTIONS ON COGNITIVE AND DEVELOPMENTAL SYSTEMS 1 Artificial Cognitive Systems that can answer Human Creativity Tests: An Approach and Two Case Studies Ana-Maria Oltet ¸eanu, Zoe Falomir, Christian Freksa Bremen Spatial Cognition Center, Universit¨ at Bremen Abstract—Creative cognitive systems are rarely assessed with the same tools as human creativity. In this paper, an approach is proposed for building cognitive systems which can solve human creativity tests. The importance of using cognitively viable processes, cognitive knowledge acquisition and organization, and cognitively comparable evaluation when implementing creative problem-solving systems is emphasized. Two case studies of arti- ficial cognitive systems evaluated with human creativity tests are reviewed. A general approach is put forward. The applicability of this general approach to other creativity tests and artificial cognitive systems, together with ways of performing cognitive knowledge acquisition for these systems are then explored. Index Terms—Computational Creativity, Human Creativity, Cognitive Systems, Cognitive Processing, Cognitive Knowledge Acquisition, Cognitively-comparable Evaluation I. I NTRODUCTION W E envision a future where cognitive agents help hu- mans solve problems in their daily routines at home and at work. In order to be helpful, those agents must find good solutions or creative alternatives the human can select from. Imagine you are in your house, needing to solve a particular problem. You need a particular tool or object (e.g. a piece of string) or recipe ingredient (e.g. mince meat) but you do not have it in the house. You would like to solve your problem using a different tool, object or ingredient, and have a system show you what you could use instead (e.g. dental floss to replace the piece of string, aubergine to replace the mince meat - depending on task and recipe context). Or imagine you are in a situation in which you ran out of ideas on how to solve a particular problem, and would like to think of a completely different approach. Let us say that you would like the help of a system that could inspire you. Such a system would need to operate or at least communicate in a cognitive manner – understand what type of information is associated with the problem at hand, in a manner that would enhance your creative process – for example by predisposing you to new ways of seeing the problem or re-representing its content. It could show you items (websites, research articles, excerpts from encyclopedias, music, films, photos, formulas) which could trigger for you new associations, new ways of solving or framing that problem. Corresponding author: Ana-Maria Oltet ¸eanu, amoodu@informatik.uni- bremen.de Manuscript received...; revised... Furthermore, such assistive systems would need to be capa- ble of creative problem-solving and of presenting their results in a manner that can easily be used as input by humans. They would need to be endowed with cognitive knowledge acquisition, and types of knowledge processing which are akin to those used by humans in their creative problem-solving. Computational creativity is a strongly emerging field in Artificial Intelligence, with application systems ranging from poetry [1], music [2], painting [3] to mathematics [4]. Em- pirical tests of human creativity and creative problem-solving do exist [5], [6], [7], [8], [9], [10]. However, computational creativity systems can rarely be assessed in a comparable manner, i.e. by using human creativity tests. This has as consequence the fact that not many artificial cognitive systems which can be used as cognitive models exist (e.g. [11]), and thus not as much progress is made as possible in terms of understanding the cognitive bases of the creative process, and in building artificial cognitive creative systems. This paper argues that (i) artificial cognitive systems can be used to shed light on the human creative process and (ii) knowledge obtained from creativity tests can be used to inform and evaluate artificial cognitive systems, if more work is done on artificial cognitive systems that can be assessed with human creativity tests. To support this claim, this paper presents a general approach for building artificial cognitive systems that can solve human creativity tests in a cognitive manner. The types of knowledge, knowledge acquisition, cognitive processes and cognitive evaluation which can be used are also discussed. The benefits of bridging this gap are then shown in terms of (a) new relations which can be observed or studied from a cognitive modeling paradigm and (b) knowledge and data obtained from human creativity tests which can then be used to inform artificial cognitive systems. Two case studies of systems which can solve human creativity tests are then presented, briefly describing how cognitive knowledge was acquired and organized for such systems, the processes used and how the systems were evaluated compared to their human counterparts. The rest of this paper is structured as follows. Section II describes the differences between computational creativity evaluation methods and human creative evaluation. Cognitive processes relevant when implementing artificial creative cog- nitive systems are discussed in Section III. Sections IV-A and IV-B briefly describe two case studies of artificial cognitive systems which can give comparable results to humans in cre-