arXiv 机器学习· Marvin van Aalst (Computational Life Science, Faculty of Mathematics and Natural Sciences, RWTH Aachen University, Germany), Ad\`ele H. Ribeiro (Causal and Explainable Artificial Intelligence in Life Sciences, Faculty of Mathematics and Natural Sciences, RWTH Aachen University, Germany, Center for Computational Life Sciences, RWTH Aachen University, Germany), Anna Matuszy\'nska (Computational Life Science, Faculty of Mathematics and Natural Sciences, RWTH Aachen University, Germany, Center for Computational Life Sciences, RWTH Aachen University, Germany)·· 5 小时前AI 评分31
当人人都能构建模型：AI 时代的机制推断
When Everyone Can Build a Model: Mechanistic Inference in the Age of AI
AI 导读
针对深度学习与 AI 辅助代码生成降低生物系统灰盒建模门槛所带来的解释难题，来自亚琛工业大学的研究探讨了机制推断的可靠性问题。研究指出预测成功并不等同于机制正确，生物建模的欠定性会导致学习组件吸收机制误差。结合因果推断可明确可识别的因果关系与干预效应，集成分析则能量化模型不确定性并指导后续验证实验。
来源：arXiv 机器学习 · arxiv.org