# Reinforced Internal-External Knowledge Synergistic Reasoning for Efficient Adaptive Search Agent

arXiv: https://arxiv.org/abs/2505.07596 · HTML: https://mlsanity.com/paper/2505.07596/

## Key facts

- Cited by 30 papers in the map (data through 14 Oct 2026).
- Placed in Knowledge- and tool-augmented reasoning, one of 17 directions in LLM reasoning.
- Builds on 1 of its 46 references that are in the map.
- First posted on arXiv 12 May 2025.

## Placed in

- LLM reasoning / [Knowledge- and tool-augmented reasoning](https://mlsanity.com/lr/knowledge-and-tool-augmented-reasoning/)
- Retrieval-augmented generation / [Agentic RAG, agents and memory](https://mlsanity.com/rag/agentic-rag-agents-and-memory/)
- LLM agents and tool use / [Web, search, and research agents](https://mlsanity.com/agents/web-search-and-research-agents/)

## Builds on (most cited first, 1 of 1)

- [Text Embeddings by Weakly-Supervised Contrastive Pre-training](https://mlsanity.com/paper/2212.03533/) (2022)

## Built on by (most cited first, 30 of 30)

- [WideSearch: Benchmarking Agentic Broad Info-Seeking](https://mlsanity.com/paper/2508.07999/) (2025)
- [A Survey of LLM-based Deep Search Agents: Paradigm, Optimization, Evaluation, and Challenges](https://mlsanity.com/paper/2508.05668/) (2025)
- [From Web Search towards Agentic Deep Research: Incentivizing Search with Reasoning Agents](https://mlsanity.com/paper/2506.18959/) (2025)
- [AutoTIR: Autonomous Tools Integrated Reasoning via Reinforcement Learning](https://mlsanity.com/paper/2507.21836/) (2025)
- [Toward Effective Tool-Integrated Reasoning via Self-Evolved Preference Learning](https://mlsanity.com/paper/2509.23285/) (2025)
- [Toward Efficient Agents: Memory, Tool learning, and Planning](https://mlsanity.com/paper/2601.14192/) (2026)
- [Reinforcement Learning Foundations for Deep Research Systems: A Survey](https://mlsanity.com/paper/2509.06733/) (2025)
- [HiPRAG: Hierarchical Process Rewards for Efficient Agentic Retrieval Augmented Generation](https://mlsanity.com/paper/2510.07794/) (2025)
- [DecEx-RAG: Boosting Agentic Retrieval-Augmented Generation with Decision and Execution Optimization via Process Supervision](https://mlsanity.com/paper/2510.05691/) (2025)
- [Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application](https://mlsanity.com/paper/2606.12191/) (2026)
- [ToolOmni: Enabling Open-World Tool Use via Agentic learning with Proactive Retrieval and Grounded Execution](https://mlsanity.com/paper/2604.13787/) (2026)
- [AdaSearch: Balancing Parametric Knowledge and Search in Large Language Models via Reinforcement Learning](https://mlsanity.com/paper/2512.16883/) (2025)
- [Learning from the Irrecoverable: Error-Localized Policy Optimization for Tool-Integrated LLM Reasoning](https://mlsanity.com/paper/2602.09598/) (2026)
- [Beyond the limitation of a single query: Train your LLM for query expansion with Reinforcement Learning](https://mlsanity.com/paper/2510.10009/) (2025)
- [ET-Agent: Incentivizing Effective Tool-Integrated Reasoning Agent via Behavior Calibration](https://mlsanity.com/paper/2601.06860/) (2026)
- [LiveSearchBench: An Automatically Constructed Benchmark for Retrieval and Reasoning over Dynamic Knowledge](https://mlsanity.com/paper/2511.01409/) (2025)
- [DocArena: Turning Raw Documents into Controllable Training Environments for Document Search Agents](https://mlsanity.com/paper/2606.26122/) (2026)
- [SAAS: Self-Aware Reinforcement Learning for Over-Search Mitigation in Agentic Search](https://mlsanity.com/paper/2605.29796/) (2026)
- [KbSD: Knowledge Boundary aware Self-Distillation for Behavioral Calibration in Agentic Search](https://mlsanity.com/paper/2606.29863/) (2026)
- [One Policy, Any Budget: Internalizing Budget-Aware Search via Reinforcement Learning](https://mlsanity.com/paper/2609.00813/) (2026)
- [RAGShaper: Eliciting Sophisticated Agentic RAG Skills via Automated Data Synthesis](https://mlsanity.com/paper/2601.08699/) (2026)
- [Mitigating Context Interference for Reliable and Efficient Search Agents](https://mlsanity.com/paper/2608.10743/) (2026)
- [SubSearch: Intermediate Rewards for Unsupervised Guided Reasoning in Complex Retrieval](https://mlsanity.com/paper/2604.07415/) (2026)
- [HindSearch: Trajectory-Level Hindsight Critique for Search-Augmented Reinforcement Learning](https://mlsanity.com/paper/2608.01597/) (2026)
- [CRISP: Critical Step Perception for Training Efficient Deep Search Agents](https://mlsanity.com/paper/2608.01867/) (2026)
- [CoBRA: Learning Tool-Use Boundaries via Counterfactual Margins](https://mlsanity.com/paper/2609.00967/) (2026)
- [SearchAtlas: Analyzing Agentic Search Strategies via Evidential Query Graphs](https://mlsanity.com/paper/2609.10901/) (2026)
- [Federated Agent Optimization](https://mlsanity.com/paper/2610.01195/) (2026)
- [SOAR: Source-Aware Reinforcement Learning with Precise Credit Assignment for Search Agents](https://mlsanity.com/paper/2610.13157/) (2026)
- [AgentTune: A Unified Training Lifecycle for Tool-Using LLM Agents](https://mlsanity.com/paper/2610.13335/) (2026)

## Abstract (arXiv)

Retrieval-augmented generation (RAG) is a common strategy to reduce hallucinations in Large Language Models (LLMs). While reinforcement learning (RL) can enable LLMs to act as search agents by activating retrieval capabilities, existing ones often underutilize their internal knowledge. This can lead to redundant retrievals, potential harmful knowledge conflicts, and increased inference latency. To address these limitations, an efficient and adaptive search agent capable of discerning optimal retrieval timing and synergistically integrating parametric (internal) and retrieved (external) knowledge is in urgent need. This paper introduces the Reinforced Internal-External Knowledge Synergistic Reasoning Agent (IKEA), which could indentify its own knowledge boundary and prioritize the utilization of internal knowledge, resorting to external search only when internal knowledge is deemed insufficient. This is achieved using a novel knowledge-boundary aware reward function and a knowledge-boundary aware training dataset. These are designed for internal-external knowledge synergy oriented RL, incentivizing the model to deliver accurate answers, minimize unnecessary retrievals, and encourage appropriate external searches when its own knowledge is lacking. Evaluations across multiple knowledge reasoning tasks demonstrate that IKEA significantly outperforms baseline methods, reduces retrieval frequency significantly, and exhibits robust generalization capabilities.