In open-domain multi-hop question answering (QA), LLM-based search agents provide a promising approach for knowledge-intensive QA by combining retrieval with reasoning. Existing methods primarily enhance open-domain multi-hop QA through reasoning paradigms, retrieval interaction, and search strategy optimization. However, employing multiple search trajectories introduces a challenging final answer selection problem. Different trajectories may support distinct candidates, and the retrieved information can be heterogeneous, redundant, incomplete, or conflicting. Directly comparing raw trajectories exposes the verifier to noisy and unaligned content, while comparing answer strings overlooks the supporting evidence for each candidate, complicating reliable final selection. To tackle this issue, we propose STEC, an evidence compression framework for final answer selection in multi-hop QA. STEC selects the final answer from the existing candidate set through two mechanisms: (1) Answer-Level Evidence Compression, which groups trajectories by normalized answer identity and converts each answer group into a candidate-specific evidence representation; and (2) Evidence-Guided Answer Verification, which compares these representations and selects the final answer from the candidate set. This design shifts the focus of final selection from raw trajectory comparison to candidate-level evidence comparison. We evaluate STEC on four open-domain multi-hop QA benchmarks against representative baselines. Experimental results demonstrate that STEC performs best overall among the compared methods, and ablation results provide evidence that answer-level evidence compression contributes to final answer selection.
Blogger's Review: The STEC framework significantly enhances the accuracy and efficiency of multi-hop question answering by introducing evidence compression mechanisms. This method not only optimizes the answer selection process but also effectively addresses the issue of information heterogeneity, making it a valuable candidate for practical applications. Its design philosophy opens new research directions for future QA systems.