Retrieval as Generation: A Unified Framework with Self-Triggered Information Planning

2026-04-13

AI Summary

This work addresses the challenge of achieving end-to-end coordination between retrieval and generation in traditional retrieval-augmented generation (RAG) systems. The authors propose GRIP, a novel framework that integrates retrieval into autoregressive generation through control tokens, dynamically determining when to retrieve, how to reformulate queries, and when to terminate retrieval. Its core innovation is the Self-Triggered Information Planning mechanism, which unifies multi-step reasoning and real-time evidence integration within a single generation trajectory. GRIP employs structured supervision signals to train retrieval behaviors across diverse scenarios—including answerable, partially answerable, and multi-hop queries. Experimental results demonstrate that GRIP significantly outperforms strong RAG baselines on five question-answering benchmarks, achieving performance comparable to GPT-4o with substantially fewer parameters.

Technology Category

• Natural Language Processing: Generation
• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

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

Abstract

We revisit retrieval-augmented generation (RAG) by embedding retrieval control directly into generation. Instead of treating retrieval as an external intervention, we express retrieval decisions within token-level decoding, enabling end-to-end coordination without additional controllers or classifiers. Under the paradigm of Retrieval as Generation, we propose \textbf{GRIP} (\textbf{G}eneration-guided \textbf{R}etrieval with \textbf{I}nformation \textbf{P}lanning), a unified framework in which the model regulates retrieval behavior through control-token emission. Central to GRIP is \textit{Self-Triggered Information Planning}, which allows the model to decide when to retrieve, how to reformulate queries, and when to terminate, all within a single autoregressive trajectory. This design tightly couples retrieval and reasoning and supports dynamic multi-step inference with on-the-fly evidence integration. To supervise these behaviors, we construct a structured training set covering answerable, partially answerable, and multi-hop queries, each aligned with specific token patterns. Experiments on five QA benchmarks show that GRIP surpasses strong RAG baselines and is competitive with GPT-4o while using substantially fewer parameters.

Problem

• Retrieval-Augmented Generation
• Information Planning
• Autoregressive Generation
• Dynamic Retrieval
• Multi-hop Reasoning

Innovation

• Retrieval as Generation
• Self-Triggered Information Planning
• Generation-guided Retrieval
• Control Tokens
• End-to-End RAG