SMADE-IE: Sparse Multi-Agent Framework with Evidence-Driven Debate for Zero-Shot Information Extraction
2026-06-03 • Computation and Language
Computation and Language
AI summaryⓘ
The authors introduce SMADE-IE, a new way to extract information using large language models without needing extra training. Their system smartly decides if it should look at the whole input at once or focus on specific types, which helps avoid errors and confusion. When predictions conflict, SMADE-IE uses a structured debate method that weighs evidence to pick the best answer. Tests on multiple datasets show it works better and uses fewer resources than previous zero-shot methods.
zero-shot information extractionlarge language modelsmulti-agent systemspromptingadaptive mode selectorevidence-driven debateToulmin argument modelnamed entity recognition (NER)relation extraction (RE)joint extraction (JERE)
Authors
Kenfeng Huang, Yi Cai, Xin Wu, Zikun Deng, Li Yuan
Abstract
Zero-shot information extraction (IE) with large language models (LLMs) has attracted increasing attention due to its flexibility in adapting to new schemas and domains without task-specific training. Existing approaches mainly rely on monolithic prompting, each-type prompting, or multi-agent debate. However, monolithic prompting often suffers from boundary and type errors, while each-type prompting and multi-agent debate introduce cross-type conflicts, redundant agent interactions, and substantial token overhead. To address these challenges, we propose SMADE-IE, a sparse and evidence-driven multi-agent framework for zero-shot IE. SMADE-IE first employs an Adaptive Mode Selector to dynamically route inputs into either a lightweight Global Extraction Mode or a Type-Centric Extraction Mode, reducing unnecessary type selection and reasoning noise. For conflicting predictions, we further introduce an Evidence-Driven Debate mechanism that structures arguments into Toulmin-style components and performs confidence aggregation through external evidence scoring and Bayesian updates. Experimental results on 9 benchmark datasets across NER, RE, and JERE tasks show that SMADE-IE consistently outperforms existing zero-shot IE baselines while also improving token efficiency through sparse agent selection and early-stopping debate.