UnfoldX, a bilingual lab for AI that studies the real world

Field notes

The lab’s public structure links methods to application domains rather than presenting AI as a general-purpose promise. Its three tracks cover multimodal, multi-view, and graph learning; gene–environment–phenotype analysis and virtual cells; and renewable-energy trading and operational optimization. The homepage also records collaborations on RNA structure prediction and a stated deployment of a power-trading forecasting model at a national energy-storage center.