Metacognitive Reasoning in Energy Based Models using Instance Based Learning Theory

2026-09-30

AI Summary

This study addresses the metacognitive reasoning challenge in large language models (LLMs) arising from their lack of uncertainty quantification and dynamic computational allocation capabilities. To this end, we propose MERITED, a novel framework that pioneers the integration of Instance-Based Learning Theory (IBLT) into Energy-Based Models (EBMs). By leveraging a dynamic computational resource scheduling algorithm, the framework achieves uncertainty-driven metacognitive control and adaptive compute allocation. Furthermore, we release an open-source 191-million-parameter reasoning model that enables efficient and interpretable uncertainty modeling. Experimental results demonstrate that our approach significantly enhances computational resource utilization efficiency during the LLM inference phase, offering a principled solution for cost-aware and uncertainty-calibrated reasoning.

Abstract

Metacognition involves reasoning about cognitive processes themselves. An example is in resource allocation where we choose how much time and effort to put into a reasoning task before we begin based on our confidence. Current Artificial Intelligence (AI) systems that rely on Large Language Models (LLMs) cannot estimate their uncertainty about an output without first responding, and cannot dynamically allocate resources to producing an output, making this type of metacognitive process difficult. A recently proposed alternative to classic transformer architectures that addresses these two concerns is the Energy Based Model (EBM) which allows for interpretable uncertainty modeling and dynamic allocation of compute resources. While EBMs can allow for control of these two processes, the actual metacognitive task of determining compute allocation based on uncertainty is not directly addressed. Instance-Based Learning Theory (IBLT) provides an approach to modeling human-like decisions from experience that has previously been applied to predicting human metacognitive reasoning. In this paper we introduce a framework for MEtacognitive Reasoning with Instance-based Learning Theory and Energy Dynamics (MERITED). Grounded in IBLT, this framework allows for control of the computational effort allocated in an EBM to allow for metacognitive control over reasoning effort based on uncertainty while remaining computationally efficient. This work has two main contributions, the training and open weight sharing of a 191M parameter reasoning EBM, and an implementation of the MERITED framework for dynamic compute allocation using an IBL model.

Problem

• Metacognitive Reasoning
• Energy Based Models
• Uncertainty Estimation
• Dynamic Compute Allocation
• Instance-Based Learning Theory

Innovation

• Metacognitive Reasoning
• Energy Based Models
• Instance-Based Learning Theory
• Dynamic Compute Allocation
• Uncertainty Modeling