Agent-centric learning: from external reward maximization to internal knowledge curation
This work addresses the issue of adaptability in general intelligence for AI researchers, proposing a foundational shift rather than an incremental improvement.
The paper tackles the problem of specialized agents lacking adaptability by proposing representational empowerment, a new agent-centric learning paradigm that focuses on controllably maintaining and diversifying internal knowledge structures, with the result being a shift towards designing more adaptable intelligent systems.
The pursuit of general intelligence has traditionally centered on external objectives: an agent's control over its environments or mastery of specific tasks. This external focus, however, can produce specialized agents that lack adaptability. We propose representational empowerment, a new perspective towards a truly agent-centric learning paradigm by moving the locus of control inward. This objective measures an agent's ability to controllably maintain and diversify its own knowledge structures. We posit that the capacity -- to shape one's own understanding -- is an element for achieving better ``preparedness'' distinct from direct environmental influence. Focusing on internal representations as the main substrate for computing empowerment offers a new lens through which to design adaptable intelligent systems.