To Know is to Construct: Schema-Constrained Generation for Agent Memory

2026-04-21

AI Summary

This work addresses two key limitations in existing agent memory systems: dense retrieval often fails to distinguish semantically similar yet contextually distinct memory items, introducing noise, while open-ended generation is prone to structural hallucinations. Inspired by constructivist epistemology, this study introduces cognitive schemata into large language model memory mechanisms for the first time, proposing a schema-constrained generation approach that ensures only valid memory keys are produced. By integrating assimilation and accommodation processes, the method enables long-term memory evolution. Dynamic maintenance of cognitive schemata, coupled with activation propagation over an associative graph, supports multi-hop reasoning. Evaluated on the LoCoMo benchmark, the proposed approach significantly outperforms all retrieval baselines, achieving consistent performance gains across all categories.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Constructivist epistemology argues that knowledge is actively constructed rather than passively copied. Despite the generative nature of Large Language Models (LLMs), most existing agent memory systems are still based on dense retrieval. However, dense retrieval heavily relies on semantic overlap or entity matching within sentences. Consequently, embeddings often fail to distinguish instances that are semantically similar but contextually distinct, introducing substantial noise by retrieving context-mismatched entries. Conversely, directly employing open-ended generation for memory access risks "Structural Hallucination" where the model generates memory keys that do not exist in the memory, leading to lookup failures. Inspired by this epistemology, we posit that memory is fundamentally organized by cognitive schemas, and valid recall must be a generative process performed within these schematic structures. To realize this, we propose SCG-MEM, a schema-constrained generative memory architecture. SCG-MEM reformulates memory access as Schema-Constrained Generation. By maintaining a dynamic Cognitive Schema, we strictly constrain LLM decoding to generate only valid memory entry keys, providing a formal guarantee against structural hallucinations. To support long-term adaptation, we model memory updates via assimilation (grounding inputs into existing schemas) and accommodation (expanding schemas with novel concepts). Furthermore, we construct an Associative Graph to enable multi-hop reasoning through activation propagation. Experiments on the LoCoMo benchmark show that SCG-MEM substantially improves performance across all categories over retrieval-based baselines.

Problem

• agent memory
• dense retrieval
• structural hallucination
• schema-constrained generation
• cognitive schema

Innovation

• Schema-Constrained Generation
• Constructivist Epistemology
• Structural Hallucination
• Cognitive Schema
• Associative Graph