To Retrieve or To Think? An Agentic Approach for Context Evolution

2026-01-13

AI Summary

This work proposes Agentic Context Evolution (ACE), a novel framework that addresses the high computational cost and noise introduced by existing retrieval-augmented methods, which perform retrieval at every reasoning step. ACE introduces metacognitive mechanisms into context evolution through a central coordinator agent that dynamically decides whether to retrieve new evidence or reuse existing knowledge, enabling adaptive alternation between retrieval and reasoning. The framework employs a multi-agent architecture comprising a coordinator, a retriever, and a reasoner, with strategy selection governed by a majority voting mechanism. Evaluated on multiple multi-hop question answering benchmarks, ACE significantly outperforms current approaches, achieving higher accuracy while substantially reducing token consumption—demonstrating efficient and precise knowledge-intensive reasoning.

Technology Category

• Multiagent Systems: Mechanism Design
• Knowledge Representation and Reasoning: Reasoning with Beliefs
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Current context augmentation methods, such as retrieval-augmented generation, are essential for solving knowledge-intensive reasoning tasks. However, they typically adhere to a rigid, brute-force strategy that executes retrieval at every step. This indiscriminate approach not only incurs unnecessary computational costs but also degrades performance by saturating the context with irrelevant noise. To address these limitations, we introduce Agentic Context Evolution (ACE), a framework inspired by human metacognition that dynamically determines whether to seek new evidence or reason with existing knowledge. ACE employs a central orchestrator agent to make decisions strategically via majority voting. It aims to alternate between activating a retriever agent for external retrieval and a reasoner agent for internal analysis and refinement. By eliminating redundant retrieval steps, ACE maintains a concise and evolved context. Extensive experiments on challenging multi-hop QA benchmarks demonstrate that ACE significantly outperforms competitive baselines in accuracy while achieving efficient token consumption. Our work provides valuable insights into advancing context-evolved generation for complex, knowledge-intensive tasks.

Problem

• retrieval-augmented generation
• context augmentation
• knowledge-intensive reasoning
• redundant retrieval
• context noise

Innovation

• Agentic Context Evolution
• retrieval-augmented generation
• metacognition
• context evolution
• multi-hop QA