Concept Formation in the Ouroboros Model

According to the Ouroboros Model several occasions can be distinguished over the course of the general autonomous cyclic activity in which new concepts are established and associated memories are preferentially laid down. Whereas a rather standard habituation process can lead to the extraction of statistical regularities from repeated common (consecutive) excitation, two specific instances are peculiar to the Ouroboros Model; the consumption analysis process marks events when especially successful or the contrary. In addition, new concepts can be assembled very quickly by combining previously existing building blocks. Relations of these theoretical considerations to supporting recent experimental findings are briefly outlined. The Ouroboros Model in a Nutshell The Ouroboros Model proposes a novel algorithmic architecture for efficient data processing in living brains and for artificial agents [1]. At its core lies a general repetitive loop where one iteration cycle sets the stage for the next. All concepts of an agent are claimed to be organized into hierarchical data structures called schemata, also known as frames, scripts or the like. During perception, any sensory input activates schemata with equal or similar constituents encountered before, thus expectations for the whole unit and especially for its parts are kindled. This corresponds to the highlighting of empty slots in the selected schema as biasing anticipated features facilitates their actual excitation. These predictions are subsequently compared to the factually encountered input. Depending on the outcome of this "consumption analysis" different next steps are taken. In case of partial fit search for further data continues; a reset, i.e. a new attempt employing another schema, is triggered if the occurring discrepancies are too big. Basically the same process applies for all other actions like memory search, active movements of the agent or language production. Self-referential monitoring of the whole process, and in particular of the flow of activation directed by the consumption analysis, yields valuable feedback for the optimum allocation of attention and resources including the selective establishment of useful new concepts. According to the Ouroboros Model basically four different situations in which novel concepts are formed and corresponding fresh memory entries are first created and shaped can be distinguished. Ways to Concept Formation Two types of occasions are directly marked in the Ouroboros Model as interesting by the outcome of the consumption analysis, and preferentially for them new records are laid down:  Events, when everything fits perfectly; i.e. associated neural representations are stored as kind of snapshots of all concurrent activity, making them available for guidance in the future as they have proved useful once.  Constellations, which led to an impasse, are worthwhile remembering, too; in this case for future avoidance. These new memories stand for junks, i.e. concepts, again as schemata, frames or scripts. Their building blocks include whatever representations are active at the time when the “snapshot” is taken, including sensory signals, abstractions, previously laid down concepts, and emotions. They might but need not include / correspond to a direct representation unit like a word. At later occasions they will serve for controlling behavior, by guiding action to or away from the marked tracks. Knowledge thus forms the very basis for the data processing steps, and its meaningful expansion is a prime outcome of its use as well; the available data base of concepts / schemata is steadily enlarged and completed, especially in areas where the need for this surfaced and is