robotic affordance learning

Compiled with AI assistance and selectively reviewed. Please consult the cited passages; the books remain the authoritative source. About this companion

Editorial synthesis · provisional

Definition

Robotic affordance learning organizes machine perception around relations among objects, possible actions and expected effects. RB2 describes learning through interaction, encoded prior knowledge and simulated reinforcement learning as distinct implementations. Such representations can support adaptation and transfer, but the source explicitly distinguishes representing value or being reward-weighted from having something matter through a system’s own vulnerable existence.

Explanation

Action-relevant features such as a graspable edge reduce the practical decision space compared with undifferentiated visual description. RB2’s quoted survey formulation treats affordances as intermediaries supporting tractable reasoning and generalization. The entry retains that engineering function without claiming a machine thereby acquires the same experience as an animal.

Some systems physically push or grasp to learn effects; others receive encoded knowledge about functions. Simulated reinforcement-learning examples provide another implementation context. The proposal keeps these routes separate rather than crediting every system with autonomous real-world discovery, an identical learning mechanism or experimentally established transfer to every new setting.

The explicit qualification distinguishes represented value, value-weighted behavior and stakes that belong to the system. Homeostatic and interoceptive machines are proposals for addressing the third achievement, not its demonstration in current affordance learners. Useful competence can therefore be acknowledged without inferring felt experience, moral standing or independent ethical answerability.

Scope and attribution

Keep technical competence and the source’s explicit homeostasis qualification without claiming machine feeling.

RB2 — The Real Book, Volume II: Becoming Is Being

Engineering implementations and the distinction among representation, weighting and own stakes are retained.

Occurrence locators: 74, 75, 76, 77.

Background used in the editorial explanation

Robotic affordance implementations are distinguished from generic ethical stakes and answerability.