面向高校智慧教育与学生成长支持场景，【研心】教育大模型集学业分析、心理评估、风险预警与智能建议于一体。For smart education and student growth support in higher education, the Yanxin educational large model integrates academic analysis, psychological assessment, risk warning, and intelligent recommendations.
系统以扩散语言模型架构为核心，围绕数据清洗、样本构建、持续预训练、监督微调、在线推理和结果展示。Centered on a diffusion language model architecture with continual pre-training, supervised fine-tuning, online inference, and result presentation.
项目综合利用学生历史成绩、课程修读信息及国民心理健康量表作答等多源数据，实现对焦虑、自闭、症状反刍因子、强迫思考、反省深思等心理指标的分析，并对下学期课程风险等级进行分层预测，自动生成结构化预警结果与针对性建议。Multi-source data—including students’ academic records, course enrollment information, and responses to national mental health scales—to analyze psychological indicators.
可为高校开展学业帮扶、心理预警、精准育人和教育治理提供数据驱动的智能支撑，体现了人工智能大模型在教育场景中的实用价值与推广潜力。 Providing data-driven intelligent support for universities in academic assistance, psychological early warning, precision education, and educational governance.
