Reinforcement Learning Measurement Model (RLMM)

Research Method · All research methods
Evidence · 6 claims (6 for) · 1 study (1 design), q2 · 1 of 1 report an effect size · 6 claims rest on one study

Description

The RLMM is a psychometric measurement framework for sequential process data that treats a person's inverse-temperature parameter βj as an ability to choose higher-valued actions under a Boltzmann policy. It replaces the MDP-MM's person-specific tabular value functions with "a shared parametric action-value function Qθ(s,a)", using centered and globally normalized advantages, a soft Bellman-consistency penalty encoding known task dynamics, and a block-coordinate MAP estimation procedure. It also yields step-level influence diagnostics that identify behaviorally critical decisions.

Accounts

• Reinforcement Learning Measurement Model (RLMM): The RLMM is a psychometric measurement framework for sequential process data that treats a person's inverse-temperature parameter βj as an ability to choose higher-valued actions under a Boltzmann policy. It replaces the MDP-MM's person-specific tabular value functions with "a shared parametric action-value function Qθ(s,a)", using centered and globally normalized advantages, a soft Bellman-consistency penalty encoding known task dynamics, and a block-coordinate MAP estimation procedure. It also yields step-level influence diagnostics that identify behaviorally critical decisions. (Wenqian Xu (2026))

Claims

• In peg-solitaire simulations with known true abilities, the RLMM attains lower RMSE of log βj than the MDP-MM on all four original boards [+M]
• RLMM runtime stays within approximately 15–20 seconds across the four peg-solitaire boards while MDP-MM runtime grows with task complexity, reaching a 19.1× speedup on the diamond board [+M]
• The RLMM recovers ability parameters on peg-solitaire boards with up to more than 23 million states, where tabular MDP-MM estimation is impractical [+M]
• In AQUALAB gameplay logs, the estimated RLMM person parameter is positively associated with cumulative trajectory reward (Pearson 0.5842, Spearman 0.5760) [+M]
• Influence-function diagnostics identify steps 3–8 of the Big-L peg-solitaire board as the most consequential region, with the lowest-β participants showing substantially larger absolute influence values there [+W]
• In AQUALAB, estimated β is ordered across episode completion status (means 1.181, 1.506, 2.343; Kruskal–Wallis p = 8.62×10−29) and higher-β quintiles complete more content more efficiently [+W]

Related Research Methods

-

Key Sources

• Wenqian Xu, Feng Ji. (2026). Reinforcement Learning Measurement Model. arXiv. https://arxiv.org/abs/2605.09305

Page metadata

Field	Value
type	research-method
id	reinforcement-learning-measurement-model-rlmm
title	Reinforcement Learning Measurement Model (RLMM)
description	The RLMM is a psychometric measurement framework for sequential process data that treats a person's inverse-temperature parameter βj as an ability to choose higher-valued actions under a Boltzmann policy.
status	review
generated	by process:settle-candidates on 2026-10-11
trust	unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.