SINAI at eRisk@CLEF 2025：基于Transformer和对话策略的抑郁症检测
SINAI at eRisk@CLEF 2025: Transformer-Based and Conversational Strategies for Depression Detection
AI总结：
本研究提出结合Transformer模型与对话策略的抑郁症检测方法，在eRisk@CLEF 2025任务2中排名第8但预测最快，在试点任务中获5队第1，验证了LLM在心理健康评估中的有效性。
摘要：
本文描述了SINAI-UJA团队参加eRisk@CLEF 2025实验室的情况。具体而言，我们处理了其中两个任务：（i）任务2：情境化抑郁症早期检测，以及（ii）试点任务：通过LLM进行对话式抑郁症检测。我们在任务2中的方法结合了广泛的预处理流程与多个基于Transformer的模型（如RoBERTa Base或MentalRoBERTa Large），以捕捉多用户对话的上下文和序列特性。对于试点任务，我们设计了一套对话策略，与基于LLM的角色进行交互，重点是在有限的对话轮次内最大化信息增益。在任务2中，我们的系统在12个参赛团队中基于F1分数排名第8。然而，更深入的分析显示，我们的模型在发出早期预测方面属于最快的之一，这在现实部署场景中是一个关键因素。这凸显了早期检测与分类准确性之间的权衡，表明未来工作中存在联合优化两者的潜在途径。在试点任务中，我们在5个团队中获得了第1名，在所有评估指标（DCHR、ADODL和ASHR）上取得了最佳整体性能。我们在此任务中的成功证明了结构化对话设计与强大语言模型相结合的有效性，强化了在敏感心理健康评估情境中部署LLM的可行性。
英文摘要：
This paper describes the participation of the SINAI-UJA team in the eRisk@CLEF 2025 lab. Specifically, we addressed two of the proposed tasks: (i) Task 2: Contextualized Early Detection of Depression, and (ii) Pilot Task: Conversational Depression Detection via LLMs. Our approach for Task 2 combines an extensive preprocessing pipeline with the use of several transformer-based models, such as RoBERTa Base or MentalRoBERTA Large, to capture the contextual and sequential nature of multi-user conversations. For the Pilot Task, we designed a set of conversational strategies to interact with LLM-powered personas, focusing on maximizing information gain within a limited number of dialogue turns. In Task 2, our system ranked 8th out of 12 participating teams based on F1 score. However, a deeper analysis revealed that our models were among the fastest in issuing early predictions, which is a critical factor in real-world deployment scenarios. This highlights the trade-off between early detection and classification accuracy, suggesting potential avenues for optimizing both jointly in future work. In the Pilot Task, we achieved 1st place out of 5 teams, obtaining the best overall performance across all evaluation metrics: DCHR, ADODL and ASHR. Our success in this task demonstrates the effectiveness of structured conversational design when combined with powerful language models, reinforcing the feasibility of deploying LLMs in sensitive mental health assessment contexts.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
