Universal Memory Architectures for Autonomous Machines
This work addresses the challenge of scalable and efficient autonomous learning for agents in unknown environments, representing a novel method rather than an incremental improvement.
The authors tackled the problem of designing a memory architecture for autonomous machines that can learn and solve problems without prior environmental knowledge, achieving provable quadratic bounds on space and time complexity while enabling minimal internal representations and recovery of the state space's homotopy type.
We propose a self-organizing memory architecture for perceptual experience, capable of supporting autonomous learning and goal-directed problem solving in the absence of any prior information about the agent's environment. The architecture is simple enough to ensure (1) a quadratic bound (in the number of available sensors) on space requirements, and (2) a quadratic bound on the time-complexity of the update-execute cycle. At the same time, it is sufficiently complex to provide the agent with an internal representation which is (3) minimal among all representations of its class which account for every sensory equivalence class subject to the agent's belief state; (4) capable, in principle, of recovering the homotopy type of the system's state space; (5) learnable with arbitrary precision through a random application of the available actions. The provable properties of an effectively trained memory structure exploit a duality between weak poc sets -- a symbolic (discrete) representation of subset nesting relations -- and non-positively curved cubical complexes, whose rich convexity theory underlies the planning cycle of the proposed architecture.