papers

Publications (17)

stat.ML2021

Revisiting Factorizing Aggregated Posterior in Learning Disentangled Representations

Ze Cheng, Juncheng Li, Chenxu Wang +4

In the problem of learning disentangled representations, one of the promising methods is to factorize aggregated posterior by penalizing the total correlation of sampled latent var…

math.AP2014

Shooting Method with Sign-Changing Nonlinearity

Ze Cheng, Congming Li

In this paper, we study the existence of solution to a nonlinear system: \begin{align} \left\{\begin{array}{cl} -Δu_{i} = f_{i}(u) & \text{in } \mathbb{R}^n, u_{i} > 0 & \text{in…

cs.LG2023

NUNO: A General Framework for Learning Parametric PDEs with Non-Uniform Data

Songming Liu, Zhongkai Hao, Chengyang Ying +3

The neural operator has emerged as a powerful tool in learning mappings between function spaces in PDEs. However, when faced with real-world physical data, which are often highly n…

cs.LG2025

Accelerating PDE-Constrained Optimization by the Derivative of Neural Operators

Ze Cheng, Zhuoyu Li, Xiaoqiang Wang +4

PDE-Constrained Optimization (PDECO) problems can be accelerated significantly by employing gradient-based methods with surrogate models like neural operators compared to tradition…

cs.LG2023

GNOT: A General Neural Operator Transformer for Operator Learning

Zhongkai Hao, Zhengyi Wang, Hang Su +6

Learning partial differential equations' (PDEs) solution operators is an essential problem in machine learning. However, there are several challenges for learning operators in prac…

cs.LG2023

A Unified Hard-Constraint Framework for Solving Geometrically Complex PDEs

Songming Liu, Zhongkai Hao, Chengyang Ying +3

We present a unified hard-constraint framework for solving geometrically complex PDEs with neural networks, where the most commonly used Dirichlet, Neumann, and Robin boundary cond…

math.AP2013

An Extended Discrete Hardy-Littlewood-Sobolev Inequality

Ze Cheng, Congming Li

Hardy-Littlewood-Sobolev (HLS) Inequality fails in the "critical" case: μ=n. However, for discrete HLS, we can derive a finite form of HLS inequality with logarithm correction for…

cs.LG2024

Reference Neural Operators: Learning the Smooth Dependence of Solutions of PDEs on Geometric Deformations

Ze Cheng, Zhongkai Hao, Xiaoqiang Wang +6

For partial differential equations on domains of arbitrary shapes, existing works of neural operators attempt to learn a mapping from geometries to solutions. It often requires a l…

cs.AI2026

BridgeDrive: Diffusion Bridge Policy for Closed-Loop Trajectory Planning in Autonomous Driving

Shu Liu, Wenlin Chen, Weihao Li +7

Diffusion-based planners have shown strong potential for autonomous driving by capturing multi-modal driving behaviors. A key challenge is how to effectively guide these models for…

cs.LG2026

Operator Learning with Domain Decomposition for Geometry Generalization in PDE Solving

Jianing Huang, Kaixuan Zhang, Youjia Wu +1

Neural operators have become increasingly popular in solving \textit{partial differential equations} (PDEs) due to their superior capability to capture intricate mappings between f…

cond-mat.quant-gas2013

Heavily Enhanced Dynamic Stark Shift in a System of Bose Einstein Condensation of Photons

Weikang Fan, Miao Yin, Ze Cheng

The dynamic Stark shift of a high-lying atom in a system of Bose Einstein condensation (BEC) of photons is discussed within the framework of nonrelativistic quantum electrodynamics…

cond-mat.mes-hall2013

Electron transport in a ferromagnetic/normal/ferromagnetic tunnel junction based on the surface of a topological insulator

Jian-Hui Yuan, Yan Zhang, Jian-Jun Zhang +1

We theoretically study the electron transport properties in a ferromagnetic/normal/ferromagnetic tunnel junction, which is deposited on the top of a topological surface. The conduc…

math.AP2016

A Liouville theorem for subcritical Lane-Emden system

Ze Cheng, Genggeng Huang, Congming Li

In this paper, we present a necessary and sufficient condition to the Lane-Emden conjecture. This condition is an energy type of integral estimate on solutions to subcritical Lane-…

math.AP2015

On the Hardy-Littlewood-Sobolev type systems

Ze Cheng, Genggeng Huang, Congming Li

In this paper, we study some qualitative properties of Hardy-Littlewood-Sobolev type systems. The HLS type systems are categorized into three cases: critical, supercritical and sub…

math.AP2016

Qualitative analysis of three-wave interaction with periodic boundary condition

Ze Cheng, Harvey Segur

First, for 3WRI with positive wave energy, we present a regularity theorem for all spatial dimension. Second, for 3WRI with negative wave energy, we present a class of solution in…

cs.LG2023

Bi-level Physics-Informed Neural Networks for PDE Constrained Optimization using Broyden's Hypergradients

Zhongkai Hao, Chengyang Ying, Hang Su +3

Deep learning based approaches like Physics-informed neural networks (PINNs) and DeepONets have shown promise on solving PDE constrained optimization (PDECO) problems. However, exi…

astro-ph.CO2014

Bose-Einstein condensation of photons in the matter-dominated universe

Ze Cheng

In 1914, Planck introduced the concept of a white body. In nature, no true white bodies are known. We assume that the universe after last-scattering is an ideal white body that con…