Jiaqi Wang

By representing prototypes as orthonormal bases on the Stiefel manifold, this work makes prototype collapse infeasible by construction, leading to more interpretable and accurate image recognition.

By explicitly modeling operation sequences on machines with a novel heterogeneous disjunctive graph, MIStar significantly boosts the performance of improvement-based methods for flexible job-shop scheduling.

Achieve reliable uncertainty estimates in federated learning with FedWQ-CP, a surprisingly simple one-shot calibration method that handles both data and model heterogeneity.

By mimicking how humans use visual anchors, ChartVSR lets models iteratively correct their own visual perception errors, leading to more accurate chart parsing.