Cross-Scale Transfer Learning for Depression Severity Prediction: From PHQ-8 to HAMD-17 Across Languages and Clinical Paradigms

AI Digest

研究提出跨尺度迁移学习框架，通过LoRA协议实现PHQ-8到HAMD-17的抑郁严重程度预测，验证了跨语言和临床范式的模型适应性。

This study proposes a cross-scale transfer learning framework using LoRA to predict depression severity from PHQ-8 to HAMD-17 across languages, demonstrating effective model adaptation.

Key points

• 提出顺序低秩适应协议实现跨尺度迁移学习 Proposes sequential low-rank adaptation for cross-scale transfer learning
• 对比不同模型规模在稀缺数据下的预测性能 Compares model performance on data-scarce target tasks
• 验证中文原生输入优于机器翻译效果 Demonstrates native Chinese input superiority over translation
• 消融实验揭示源监督对齐的关键作用 Ablation shows source supervision alignment importance

Takeaway: 跨语言迁移学习在临床抑郁评估中展现显著效果。 / Cross-lingual transfer learning shows significant efficacy in clinical depression assessment.

Why it matters 方法创新性强，实验设计严谨，对多语言医疗AI开发有实践指导价值。