On social media, several individuals experiencing suicidal ideation (SI) do not disclose their distress explicitly. Instead, signs may surface indirectly through everyday posts or peer interactions. Detecting such implicit signals early is critical but remains challenging. We frame early and implicit SI as a forward-looking prediction task and develop a computational framework that models a user's information environment, consisting of both their longitudinal posting histories as well as the discourse of their socially proximal peers. We adopted a composite network centrality measure to identify top neighbors of a user, and temporally aligned the user's and neighbors' interactions -- integrating the multi-layered signals in a fine-tuned DeBERTa-v3 model. In a Reddit study of 1,000 (500 Case and 500 Control) users, our approach improves early and implicit SI detection by an average of 10% over all other baselines. These findings highlight that peer interactions offer valuable predictive signals and carry broader implications for designing early detection systems that capture indirect as well as masked expressions of risk in online environments.

翻译：在社交媒体上，部分经历自杀意念（SI）的个体不会明确表露其心理困扰。相反，迹象可能通过日常发帖或同伴互动间接显现。及早检测此类隐性信号至关重要，但仍具挑战性。我们将早期隐性自杀意念定义为一个前瞻性预测任务，并开发了一个计算框架，该框架对用户的信息环境进行建模，包括其纵向发帖历史及其社交邻近同伴的对话内容。我们采用复合网络中心性度量来识别用户的顶部邻居，并在时间维度上对齐用户与邻居的互动——通过微调的DeBERTa-v3模型整合多层信号。在一项针对1,000名（500名案例组与500名对照组）Reddit用户的研究中，我们的方法在所有基线模型上将早期隐性自杀意念检测的平均性能提升了10%。这些发现表明，同伴互动提供了有价值的预测信号，并对设计能够捕捉在线环境中间接及隐匿风险表达的早期检测系统具有更广泛的启示意义。