Details
Status: Available
Type(s): Master Thesis
Participants: 1 Student
Keywords: Neural Networks, Biologically Plausible Neurolearning, Artificial Consciousness, AI
Task Description & Requirements
Based on the given body of literature develop a simple model of artificial consciousness (AC) and implement it in a high-level programming language (either Java or Python). The functionality must include (1) a mechanism for perception + an inner monitor of such perception, (2) an idea generation and monitoring component (= includes the memory), (3) a filtering mechanism that separates “healthy” from other ideas and (4) leads to behavior. Ideally, the AC mechanism should be implemented for a toy scenario of, for example, an insect living in a terrarium (i.e. a simplified world) and interacting with its environment. If possible, the model should be designed for extension so that more and more complex perception, reasoning and behavioral components can be added. The model should be implemented in two steps: (a) as a static solution for AC problems and (b) as an evolutionary model that learns and develops further. In the present work, the focus is on task a), as task b) is clearly beyond the current research frontier. Students of the courses Media and Brain 1/2 are preferred for this task.