اتصال فضاهای بازنمودی از رنگ ها و احساسات برای خلاقیت

To implement cognitive functions such as creativity, or the ability to create analogies and metaphors, it is important to have mechanisms binding different representational spaces. The paper discusses this issue in the broader context of having a “artist” robot, able to process his visual perception, to use his experience and skills as a painter, and to develop a creative digital artefact. In this context, two different spaces of color representation are respectively used to associate a linguistic label and an emotional value to color palettes. If the goal is to build an image that communicates a desired emotion, the robot can use a neural architecture to choose the most suitable palette. The experience concerning palette-emotion association is derived from the analysis of data enriched with textual description available on the web. The representation of colors and palettes is obtained by using neural networks and self association mechanisms with the aim of supporting the choice of the palette.

The presented work takes place in the generic context of creativity and aims at investigating a basic mechanism that can be used for the construction of new representational spaces. A much-discussed theory was originated by the so-called model of blending (or conceptual integration) which identifies and connects mental spaces among them ( Fauconnier & Turner, 1998). Considering cognitive mechanisms for creativity, three components generate a blend: the composition (or fusion) which pairs elements from the input spaces into the blend; the completion (or emergence) of a pattern in the blend, which is filled using long-term memory information; the process that simulate a cognitive work and its performance evaluation ( [Abdel-Fattah et al., 2012] and [Pereira, 2007]). In our opinion it is possible to have robust fusion algorithms and completion through the combination of various models of neural networks: an example of such an approach is described in Thagard and Stewart, 2011 that makes it possible to emphasize associations useful to generate creative ideas by simple vector convolution. The importance of associative mechanisms is also underlined by neurobiological models of creativity, many of which are based on the simultaneous activation and communication between brain regions that are generally not strongly connected ( Heilman, Nadeau, & Beversdorf, 2003). We propose an architecture that produces links between representational spaces originated from visual perception. In particular, we address two interesting cognitive aspects: the association between the name and the perception of color and the association between palette of colors and emotional labels. The two representation spaces are implemented by using neural networks, and from them new bindings in a “creative” space can arise. Extending a previous work (Infantino, Pilato, Rizzo, & Vella, 2013), we choose to put together a raw color representation with a link between emotion and color sets using some recent hypotheses that try to explain the mechanisms of creativity (Boden, 2009). Currently we have built a subsystem able to connect emotions and color palettes. This connection is aimed at obtaining a suitable palette if a specific emotion is required, in an image or a graphic artefact. This module is a part of a larger project that is focused on obtaining a robotic system able to paint and reproduce a portrait of a human subject. The project is aimed at combining an approach based on cognitive architectures ([Goertzel et al., 2010] and [Langley et al., 2009]) with mechanisms of computational creativity ([Boden, 2009] and [Colton et al., 2009]) trying to get an implementation that satisfies both software and hardware constraints, which are introduced when working with a real robotic platform. All cognitive architectures presented in the literature do not explicitly provide mechanisms to achieve creativity, partly because only recent researches are addressing some important and closely related cognitive aspects such as emotions, intentions, awareness, consciousness. These aspects, and other high level functions of new cognitive architectures are reported in Samsonovich, 2012. Given the proposed models of creativity in literature, it is possible to identify some features that are likely to produce a digital portrait painter using a robot that is able to develop his artificial visual perception and to originate a creative act based on its experience, its expertise and also through interaction with the human (both during learning and in final evaluation). The paper describes the first phase of our research project, and it aims at creating cognitive software infrastructure that provides creative skills mainly from visual perception. In a second phase, we plan to introduce the physicalness of the humanoid by including other sensory information such as auditory and tactile, to achieve perhaps the use of a real brush and a canvas. The paper is structured as follows: the first Section 2 gives a brief introduction to colors and emotion interaction as reported in relevant literature, the second Section 3 describes the proposed system, the third Section 4 illustrates the dataset used, the fourth Section 5 presents the results and finally some conclusions are drawn.

The presented work shows a feasible way to bind two different representational spaces based on the color perception. A robust architecture based on neural networks has been used to associate a linguistic label and an emotional value to color palettes. This subsystem, suitably inserted in a cognitive architecture, can be used to support the creative skills of a robot that aims at simulating the creative process of a human painter. Simultaneously with these results, we are developing another subsystem, that starting from a set of painting styles, derived from a set of famous paintings, creates his own style, combining simple filters to be applied to the scene perceived. This way, the painting created will have a defined style and a use of color that is motivated by emotions according to what the artificial artist intends to communicate. In the future we plan to check the paintings and their color palettes using human subjects in order to assess the obtained results.