15 July 2025 in Cognition
How automatic is reinforcement learning (RL)? Here, using a recent computational framework that separates contributions from working memory versus RL during instrumental learning, we asked if taxing higher executive functions influences a putatively lower-level, procedural RL system. Across three experiments, we found that dual-tasking could indeed disrupt RL, even when isolating RL from working memory's contributions to behavior. These results speak to methodological considerations in the use of dual tasks during learning, suggesting that cognitive load can interfere with multiple learning and memory systems simultaneously. Moreover, our results point to a less constrained conception of RL as a putatively low-level procedural system, supporting a view that tight links exist between executive function and subcortical learning processes.
14 August 2025 in Cognitive Science
A hallmark of effective teaching is that it grants learners not just a collection of facts about the world, but also a toolkit of abstractions that can be applied to solve new problems. How do humans teach abstractions from examples? Here, we applied Bayesian models of pedagogy to a necklace‐building task where teachers create necklaces to teach a learner “motifs” that can be flexibly recombined to create new necklaces. In Experiment 1 (N = 151), we find that human teachers produce necklaces that are simpler (i.e., have lower algorithmic complexity) than would be expected by chance, as indexed by a model that samples uniformly from all necklaces that contain the target motifs. This tendency to select simpler examples is captured by a pedagogical sampling model that tries to maximize the learner's belief in the true motifs by prioritizing examples that have fewer alternative interpretations. In Experiment 2 (N = 295), we find that simplicity is beneficial. Human learners recover the underlying motifs better when teachers produce simpler sequences, as predicted by the pedagogical sampling model. However, humans learn best from human teachers rather than from model‐generated examples, which suggests that human teachers have additional expectations about how learners will interpret examples that are not captured by standard models of teaching. Our work provides a principled framework to understand when and why teachers use simple examples to convey abstract knowledge.
28 August 2024 in PsyArXiv
A hallmark of effective teaching is that it grants learners not just a collection of facts about the world, but also a toolkit of abstractions that can be applied to solve new problems. How do humans transmit and acquire generalizable abstractions from examples? Here, we applied Bayesian models of pedagogy to a necklace-building task where teachers create necklaces to teach a learner ``motifs'' that can be flexibly recombined to create new necklaces. In Experiment 1 (N = 151), we find that human teachers produce necklaces that are simpler (i.e., have lower algorithmic complexity) than would be expected by chance, as indexed by a model that samples uniformly from all necklaces that contain the target motifs. This tendency to select simpler examples is partially captured by a pedagogical sampling model that tries to maximize the learner's belief in the underlying motifs. In Experiment 2 (N = 295), we find that simplicity is beneficial. Human learners recover the underlying motifs better when teachers produce simpler sequences, and they learn best from human teachers rather than from model-generated examples. Our results suggest that the computational principles that underlie effective communication and teaching may also provide a first step towards understanding the transmission of culturally-specific abstractions.
16 August 2025 in Communications Psychology
Through reward-based learning, people learn which actions generate which outcomes in which situations. What happens to human reward-based learning when outcomes are shared? Here we show that learning is impacted by inequity in the distribution of rewards (self-other) and others' identity. In three experiments, participants could learn how different actions, in response to different stimuli, generated different monetary rewards, each split between the participant and a member of a particular social group. Overall, participants learned more slowly and less successfully when they received a smaller (vs larger) share of the total reward. Stereotypes about the partner's social group additionally modulated learning rates when cognitive load was reduced, with lower perceived warmth or competence hampering learning from the partner's share. Computational modeling showed participants' learning was best explained by adapting the standard reinforcement learning model to account for stereotypes and inequity information, demonstrating that social context modulates non-social learning processes.