PPL
Xiao-Ping Zhang
Papers (5)
VEN-VL: A Visual Ensemble MoE Framework for Effective and Efficient Multi-Modal UnderstandingCausal Tongue-Tie: LLMs Can Encode Causal Direction, But Their Yes/No Outputs Fail to ExpressPHGNet: Prototype-Guided Hypergraph Construction for Heterogeneous Spatiotemporal ForecastingADMFormer: An Adaptive-Decomposition Transformer with Time-Varying Masked Spatial Attention for Traffic ForecastingVerifiable Process Rewards for Agentic Reasoning