Jiacheng Zhu

Papers

Total Citations

H-Index

About

Jiacheng Zhu is a researcher whose work bridges machine learning, stochastic processes, and functional data analysis, with a focus on modeling complex, sequential, and dynamic interactions. His most cited paper, "Recurrent Attentive Neural Process for Sequential Data" (2019, 23 citations), introduces a novel architecture that extends Attentive Neural Processes by incorporating recurrence, enabling adaptive prediction of distributions from observed context sets—a significant contribution to uncertainty-aware learning in time-series. In "Robust unsupervised learning of temporal dynamic vehicle-to-vehicle interactions" (2022, 7 citations), Zhu applies these ideas to autonomous systems, modeling real-world vehicular dynamics with robustness to noise. His work "Functional Optimal Transport: Mapping Estimation and Domain Adaptation for Functional Data" (2021, 2 citations) pioneers a theoretical framework for optimal transport in infinite-dimensional function spaces, using Hilbert-Schmidt operators to align distributions across domains. This foundational contribution opens new avenues for domain adaptation in functional data, such as biomedical signals or environmental monitoring. Zhu’s research is notable for its synthesis of probabilistic modeling, attention mechanisms, and geometric analysis, offering tools that are both mathematically rigorous and practically impactful for sequential decision-making and transfer learning.