Whom to Query for What: Adaptive Group Elicitation via Multi-Turn LLM Interactions
This work addresses the challenge of reducing uncertainty in collective assessments for survey designers and researchers, though it is incremental in combining existing techniques for adaptive elicitation.
The paper tackles the problem of adaptively selecting both questions and respondents to efficiently elicit group-level information under budget constraints, achieving a >12% relative gain in prediction accuracy on a real-world dataset with limited respondent budgets.
Eliciting information to reduce uncertainty about latent group-level properties from surveys and other collective assessments requires allocating limited questioning effort under real costs and missing data. Although large language models enable adaptive, multi-turn interactions in natural language, most existing elicitation methods optimize what to ask with a fixed respondent pool, and do not adapt respondent selection or leverage population structure when responses are partial or incomplete. To address this gap, we study adaptive group elicitation, a multi-round setting where an agent adaptively selects both questions and respondents under explicit query and participation budgets. We propose a theoretically grounded framework that combines (i) an LLM-based expected information gain objective for scoring candidate questions with (ii) heterogeneous graph neural network propagation that aggregates observed responses and participant attributes to impute missing responses and guide per-round respondent selection. This closed-loop procedure queries a small, informative subset of individuals while inferring population-level responses via structured similarity. Across three real-world opinion datasets, our method consistently improves population-level response prediction under constrained budgets, including a >12% relative gain on CES at a 10% respondent budget.