Skip to content

More accurate behavioral predictions with hybrid Bayesian-connectionist models

Aug 2026 · 0 citations
Computer Science

TL;DR

BBT outperforms traditional approaches at predicting human behavior while also revealing psychological insights, resulting in models that can both mimic Bayesian priors and capture heuristics and biases that violate simple modeling assumptions.

Abstract

Researchers must often choose between Bayesian or neural network models of behavior, two paradigms with complementary strengths and weaknesses. An ideal paradigm would facilitate testing many kinds of representations and inductive biases; Bayesian models make this easy, while neural networks do not. Similarly, an ideal paradigm would avoid over-simplifications; neural networks make this easy, while Bayesian models do not. Here, we introduce Bayesian distillation with Behavioral Tuning (BBT) as an approach to getting the best of both traditions. BBT offers a simple recipe for model building: first, a neural network is trained to mimic a Bayesian model through synthetic data, and second, the network is fine-tuned on human behavior to capture additional structure and nuance. Across four case studies in human concept learning, we find that BBT outperforms traditional approaches at predicting human behavior while also revealing psychological insights, resulting in models that can both mimic Bayesian priors and capture heuristics and biases that violate simple modeling assumptions.

View source

Similar papers

Better Behavioral Prediction, More Faithful Model Ablations? Evidence from Sequential Choice

Three findings emerge: in the restless task, neural models trained without rewards predict held-out choices better than four simple training-fitted behavioral baselines, and at some reward weights, neural networks predict better than a pooled reinforcement-learning model but have less faithful changes in choice probabi...

Han-Bo Xie · 0 citations
Open access Aug 2026

Prior-Likelihood Metamers to Distinguish Strategies Underlying Human Bayesian Behaviour in Perception

Bayesian decision theory proposes that people make statistically rational decisions by combining prior knowledge with sensory information (likelihoods). This framework successfully explains many aspects of human behaviour. However, debate persists over whether people perform precise Bayesian computations (i.e., explici...

Chin-Hsuan Sophie Lin, N. Terence, M. Garrido · 0 citations
Review Open access Aug 2026

Good practices for Bayesian inference in statistical learning.

Bayesian inference provides a unifying formalism for describing the inferences humans and other animals perform about features of the environment that change on fast timescales (moment-by-moment or trial-by-trial), as well as their learning of environmental contingencies and regularities that vary over slower time scal...

T. Kovács, Máté Lengyel · 1 citation
Review Open access Aug 2026

Making models disagree to learn how brains compute

Computational hypotheses about brain information processing can be expressed in neural network models. Neuroscientists have begun to compare such models in terms of their alignment with neural and behavioural data. The high parametric capacity of these models is essential to their ability to capture cognitive processes...

Tal Golan, H. Schütt, N. Kriegeskorte · 0 citations
Review Open access Aug 2026

Explaining success in Bayesian reasoning with the adaptive toolbox and expertise

Research on Bayesian reasoning has been shaped by two productive traditions: ecological rationality, which explains why natural frequencies facilitate inference, and nested sets accounts, which show how transparent set relations support analytic reasoning. Together, these approaches have generated a rich empirical lite...

Hannah Fenwick, Guillermo Campitelli, Matthew B. Thompson · 0 citations

Related blog posts

Microsoft Research Blog Nov 16, 2026

TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient

AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.