Evolving Self-supervised Neural Networks: Autonomous Intelligence from Evolved Self-teaching
This addresses the challenge of creating intrinsically motivated AI systems for autonomous agents, though it appears incremental as it builds on existing evolutionary and self-learning concepts.
The paper tackles the problem of developing autonomous intelligence by proposing evolving self-supervised neural networks, which combine evolution and self-learning to enable agents to teach themselves without external supervision. Experimental results show that only agents using this combined approach demonstrated intelligent behavior in a foraging task, unlike those using evolution or self-learning alone.
This paper presents a technique called evolving self-supervised neural networks - neural networks that can teach themselves, intrinsically motivated, without external supervision or reward. The proposed method presents some sort-of paradigm shift, and differs greatly from both traditional gradient-based learning and evolutionary algorithms in that it combines the metaphor of evolution and learning, more specifically self-learning, together, rather than treating these phenomena alternatively. I simulate a multi-agent system in which neural networks are used to control autonomous foraging agents with little domain knowledge. Experimental results show that only evolved self-supervised agents can demonstrate some sort of intelligent behaviour, but not evolution or self-learning alone. Indications for future work on evolving intelligence are also presented.