Abstract:Humans and animals routinely infer relations between different items or events and generalize these relations to novel combinations of items. This allows them to respond appropriately to radically novel circumstances and is fundamental to advanced cognition. However, how learning systems (including the brain) can implement the necessary inductive biases has been unclear. Here we investigated transitive inference (TI), a classic relational task paradigm in which subjects must learn a relation ( > and > ) and generalize it to new combinations of items ( > ). Through mathematical analysis, we found that a broad range of biologically relevant learning models (e.g. gradient flow or ridge regression) perform TI successfully and recapitulate signature behavioral patterns long observed in living subjects. First, we found that models with item-wise additive representations automatically encode transitive relations. Second, for more general representations, a single scalar “conjunctivity factor” determines model behavior on TI and, further, the principle of norm minimization (a standard statistical inductive bias) enables models with fixed, partly conjunctive representations to generalize transitively. Finally, neural networks in the “rich regime,” which enables representation learning and has been found to improve generalization, unexpectedly show poor generalization and anomalous behavior. We find that such networks implement a form of norm minimization (over hidden weights) that yields a local encoding mechanism lacking transitivity. Our findings show how minimal statistical learning principles give rise to a classical relational inductive bias (transitivity), explain empirically observed behaviors, and establish a formal approach to understanding the neural basis of relational abstraction.

A mediational theory of equivalence relations and transformation of function

Relating in the Wild: Toward an Analysis of Equivalence Relations Under More Naturalistic Conditions

Exploratory Study of Equivalence under Conjoint Select and Reject Control

Equivalence Projective Simulation as a Framework for Modeling Formation of Stimulus Equivalence Classes

The role of intraverbal bidirectional naming in the establishment of comparative relations

Model Equivalences

FUNCTIONAL CLASSES AND EQUIVALENCE RELATIONS

Where Hypotheses Come From: Learning New Relations by Structural Alignment

Sociocultural Theory and the Mediated Learning Experience

Distinguishing Units of Analysis, Procedures, and Processes in Cultural Selection: Notes on Metacontingency Terminology

Deriving relations at multiple levels of complexity following minimal instruction: A demonstration

Using a Respondent-Type Matching-to-Sample Exclusion Training Procedure to Establish Equivalence Responding

Abstraction, Multiple Exemplar Training and the Search for Derived Stimulus Relations in Animals

The Failings of Conventional Mediation Analysis and a Design-Based Alternative

A mathematical theory of relational generalization in transitive inference

Evaluating mechanisms of behavior change to inform and evaluate technology-based interventions

Contingency awareness in evaluative conditioning: A comment on baeyens, eelen, and van den bergh

Mediating variable

Empirically deriving omission and commission errors for relapse tests: A demonstration of reverse translation.

Relational Cues are Affectively Differentiated

Reciprocal language style matching: Indicator or facilitator of therapeutic bond