Resource-rational analysis

A cognitive-science paradigm advanced by Falk Lieder and Thomas L. Griffiths (target article, Behavioral and Brain Sciences 43, e1, 2020) that explains human cognition as the optimal use of limited computational resources. It extends Anderson's rational analysis (claim-rational-analysis-anderson-optimal-solution-marr-level) from predicting behaviour off the environment's structure to predicting cognitive mechanisms from internal resources plus the environment, pushing the program toward Marr's algorithmic level. Formally it scores a heuristic by expected utility minus an explicit opportunity-cost term on the compute it consumes, and derives the algorithm that best trades resource cost against approximation accuracy. Its "bounded" half descends from Horvitz and Russell's bounded optimality, not from Werbos's approximate dynamic programming.

References

• claim-resource-rational-analysis-extends-andersons-rational-analysis — self-described extension of Anderson, toward Marr's algorithmic level
• claim-resource-rational-analysis-descends-from-bounded-optimality-not-werbos-adp — its actual AI antecedent is bounded optimality, not Werbos's ADP
• claim-resource-rational-analysis-formalizes-utility-minus-compute-cost — the utility-minus-cost formalism (Eq 3/4)
• observation-resource-rational-analysis-confirms-anderson-leg-not-werbos-leg — confirms the Anderson leg, not the Werbos leg, of the optimization-as-theory-of-mind bridge
• Bridge context: claim-optimization-as-theory-of-mind-two-lineages