Author name cluster

Nan Ding

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

12 papers

2 author rows

AAAI Conference 2026 Conference Paper

AquaSplatting: A Hybrid 3D Representation for Robust Underwater Scene Reconstruction via Dual-Branch Rendering

Jiangbei Hu
Haobo Wang
Baixin Xu
Nan Ding
Zhimao Lu
Na Lei
Ying He

While 3D Gaussian Splatting (3DGS) excels at real-time rendering of standard scenes, it struggles to reconstruct underwater environments due to severe challenges such as light scattering, color attenuation, and sparse coverage of Gaussian kernels in far-field aqueous regions. To address this, we introduce AquaSplatting, a hybrid framework that combines explicit and implicit modeling methods for robust underwater scene reconstruction. Our dual-branch architecture employs 3DGS in a geometry-guided branch to model solid surfaces like the seabed, while a medium-aware branch uses a compact, view-dependent MLP to represent volumetric water effects. Furthermore, a neural underwater hybrid rendering mechanism adaptively fuses these two representations based on accumulated opacity. Thanks to this dual-branch framework, our method can also synthesize restored images without water medium. To enhance efficiency, our proposed engagement-based pruning (EBP) strategy quantifies each Gaussian's contribution by accumulating its image-space gradients over multiple frames, enabling the principled removal of primitives with negligible impact. The entire framework is optimized using a comprehensive loss function that integrates photometric, exposure, semantic, and depth priors to maximize visual fidelity. Experiments on challenging underwater datasets demonstrate that AquaSplatting achieves the state-of-the-art in reconstruction quality surpassing prior methods while maintaining real-time performance.

PDF Details DOI

EAAI Journal 2026 Journal Article

Collaborative deep neural network inference via distributed partitioning and resource optimization in vehicular edge computing

Yuntao Hao
Nan Ding
Huanbo Gao

In the Internet of Vehicles (IoV), collaborative inference based on Deep Neural Networks (DNNs) faces considerable challenges due to the dynamic and heterogeneous service demands of multiple vehicular devices supported by edge servers. These challenges give rise to a complex joint optimization problem that simultaneously determines the optimal DNN partition points and allocates computing resources from servers to individual devices. This problem is formulated as a Mixed Integer Nonlinear Programming (MINLP) model, involving both discrete partitioning decisions and continuous resource allocation variables, which makes it computationally demanding in large-scale and dynamic environments. To address this issue, we propose a fully distributed algorithm that achieves rapid convergence to feasible solutions. Furthermore, to accelerate the distributed iterative solving process, an adaptive strategy is designed to dynamically adjust partition points and resource allocations according to the real-time server load. For Directed Acyclic Graph (DAG)-structured DNNs, a novel structural transformation technique is developed to convert them into equivalent chain representations, enabling unified optimization across heterogeneous DNN topologies. Theoretical analysis validates the feasibility and near-optimality of the proposed approach under standard consensus assumptions, while empirical results demonstrate stable performance in dynamic vehicular environments. Extensive experiments conducted in realistic vehicular scenarios demonstrate the superior effectiveness and scalability of the proposed method in reducing inference latency and system overhead.

Details DOI

NeurIPS Conference 2021 Conference Paper

Bridging the Gap Between Practice and PAC-Bayes Theory in Few-Shot Meta-Learning

Nan Ding
Xi Chen
Tomer Levinboim
Sebastian Goodman
Radu Soricut

Despite recent advances in its theoretical understanding, there still remains a significant gap in the ability of existing PAC-Bayesian theories on meta-learning to explain performance improvements in the few-shot learning setting, where the number of training examples in the target tasks is severely limited. This gap originates from an assumption in the existing theories which supposes that the number of training examples in the observed tasks and the number of training examples in the target tasks follow the same distribution, an assumption that rarely holds in practice. By relaxing this assumption, we develop two PAC-Bayesian bounds tailored for the few-shot learning setting and show that two existing meta-learning algorithms (MAML and Reptile) can be derived from our bounds, thereby bridging the gap between practice and PAC-Bayesian theories. Furthermore, we derive a new computationally-efficient PACMAML algorithm, and show it outperforms existing meta-learning algorithms on several few-shot benchmark datasets.

PDF Details

NeurIPS Conference 2017 Conference Paper

Cold-Start Reinforcement Learning with Softmax Policy Gradient

Nan Ding
Radu Soricut

Policy-gradient approaches to reinforcement learning have two common and undesirable overhead procedures, namely warm-start training and sample variance reduction. In this paper, we describe a reinforcement learning method based on a softmax value function that requires neither of these procedures. Our method combines the advantages of policy-gradient methods with the efficiency and simplicity of maximum-likelihood approaches. We apply this new cold-start reinforcement learning method in training sequence generation models for structured output prediction problems. Empirical evidence validates this method on automatic summarization and image captioning tasks.

PDF Details

NeurIPS Conference 2016 Conference Paper

Stochastic Gradient MCMC with Stale Gradients

Changyou Chen
Nan Ding
Chunyuan Li
Yizhe Zhang
Lawrence Carin

Stochastic gradient MCMC (SG-MCMC) has played an important role in large-scale Bayesian learning, with well-developed theoretical convergence properties. In such applications of SG-MCMC, it is becoming increasingly popular to employ distributed systems, where stochastic gradients are computed based on some outdated parameters, yielding what are termed stale gradients. While stale gradients could be directly used in SG-MCMC, their impact on convergence properties has not been well studied. In this paper we develop theory to show that while the bias and MSE of an SG-MCMC algorithm depend on the staleness of stochastic gradients, its estimation variance (relative to the expected estimate, based on a prescribed number of samples) is independent of it. In a simple Bayesian distributed system with SG-MCMC, where stale gradients are computed asynchronously by a set of workers, our theory indicates a linear speedup on the decrease of estimation variance w. r. t. the number of workers. Experiments on synthetic data and deep neural networks validate our theory, demonstrating the effectiveness and scalability of SG-MCMC with stale gradients.

PDF Details

NeurIPS Conference 2015 Conference Paper

Embedding Inference for Structured Multilabel Prediction

Farzaneh Mirzazadeh
Siamak Ravanbakhsh
Nan Ding
Dale Schuurmans

A key bottleneck in structured output prediction is the need for inference during training and testing, usually requiring some form of dynamic programming. Rather than using approximate inference or tailoring a specialized inference method for a particular structure---standard responses to the scaling challenge---we propose to embed prediction constraints directly into the learned representation. By eliminating the need for explicit inference a more scalable approach to structured output prediction can be achieved, particularly at test time. We demonstrate the idea for multi-label prediction under subsumption and mutual exclusion constraints, where a relationship to maximum margin structured output prediction can be established. Experiments demonstrate that the benefits of structured output training can still be realized even after inference has been eliminated.

PDF Details

NeurIPS Conference 2015 Conference Paper

On the Convergence of Stochastic Gradient MCMC Algorithms with High-Order Integrators

Changyou Chen
Nan Ding
Lawrence Carin

Recent advances in Bayesian learning with large-scale data have witnessed emergence of stochastic gradient MCMC algorithms (SG-MCMC), such as stochastic gradient Langevin dynamics (SGLD), stochastic gradient Hamiltonian MCMC (SGHMC), and the stochastic gradient thermostat. While finite-time convergence properties of the SGLD with a 1st-order Euler integrator have recently been studied, corresponding theory for general SG-MCMCs has not been explored. In this paper we consider general SG-MCMCs with high-order integrators, and develop theory to analyze finite-time convergence properties and their asymptotic invariant measures. Our theoretical results show faster convergence rates and more accurate invariant measures for SG-MCMCs with higher-order integrators. For example, with the proposed efficient 2nd-order symmetric splitting integrator, the mean square error (MSE) of the posterior average for the SGHMC achieves an optimal convergence rate of $L^{-4/5}$ at $L$ iterations, compared to $L^{-2/3}$ for the SGHMC and SGLD with 1st-order Euler integrators. Furthermore, convergence results of decreasing-step-size SG-MCMCs are also developed, with the same convergence rates as their fixed-step-size counterparts for a specific decreasing sequence. Experiments on both synthetic and real datasets verify our theory, and show advantages of the proposed method in two large-scale real applications.

PDF Details

NeurIPS Conference 2014 Conference Paper

Bayesian Sampling Using Stochastic Gradient Thermostats

Nan Ding
Youhan Fang
Ryan Babbush
Changyou Chen
Robert Skeel
Hartmut Neven

Dynamics-based sampling methods, such as Hybrid Monte Carlo (HMC) and Langevin dynamics (LD), are commonly used to sample target distributions. Recently, such approaches have been combined with stochastic gradient techniques to increase sampling efficiency when dealing with large datasets. An outstanding problem with this approach is that the stochastic gradient introduces an unknown amount of noise which can prevent proper sampling after discretization. To remedy this problem, we show that one can leverage a small number of additional variables in order to stabilize momentum fluctuations induced by the unknown noise. Our method is inspired by the idea of a thermostat in statistical physics and is justified by a general theory.

PDF Details

ICRA Conference 2011 Conference Paper

Adaptive fuzzy control for synchronization of nonlinear teleoperators with stochastic time-varying communication delays

Zhijun Li
Nan Ding
Xiaoqing Cao

In this paper, adaptive fuzzy control is investigated for nonlinear teleoperators with time-delays, which ensures synchronization of positions and velocities of the master and slave manipulators, and does not rely on the use of the scattering transformation. Compared with previous passivity framework, the communication delays are assumed to be stochastic time varying. By feedback linearization, the nonlinear dynamics of the tele-operation system is transformed into two sub systems: local master/slave position control with unmodelled dynamics and delayed motion synchronization. Then, based on linear matrix inequalities (LMI) and Markov jump linear systems, adaptive fuzzy control strategies are developed for the nonlinear teleoperators with time-delay, modeling uncertainties and external disturbances by using the approximation property of the fuzzy logic systems. It is proven that the master-slave tele operation system is stochastically stable in mean square under specific LMI conditions, and all the signals of the resulting closed-loop system are uniformly bounded.

Details

NeurIPS Conference 2011 Conference Paper

t-divergence Based Approximate Inference

Nan Ding
Yuan Qi
S. V. N. Vishwanathan

Approximate inference is an important technique for dealing with large, intractable graphical models based on the exponential family of distributions. We extend the idea of approximate inference to the t-exponential family by defining a new t-divergence. This divergence measure is obtained via convex duality between the log-partition function of the t-exponential family and a new t-entropy. We illustrate our approach on the Bayes Point Machine with a Student's t-prior.

PDF Details

IROS Conference 2010 Conference Paper

Motion control of an autonomous vehicle based on wheeled inverted pendulum using neural-adaptive implicit control

Zhijun Li 0001
Yang Li 0029
Chenguang Yang 0001
Nan Ding

Wheeled inverted pendulum (WIP) models have been widely used in the field of autonomous robotics and intelligent vehicles. A novel transportation system, WIP-car is proposed in this paper, which is composed of a mobile wheeled inverted pendulum system, a driven chair, an acceleration pedal and a deceleration pedal, which are used to drive the chair forward or backward such that the car can be accelerated or decelerated. The neural-adaptive implicit control is designed for dynamic balance and stable tracking of desired trajectories of WIP-car. Neither the dynamics nor the dimension of the regulated system is required to be known, while the relative degree of the regulated output is assumed to be known. Under the assumption that WIP-car is feedback linearizable, adaptive neural network is introduced to cancel the inversion dynamics error. Simulation results demonstrate that the system is able to track reference signals satisfactorily with all closed loop signals uniformly bounded.

Details

NeurIPS Conference 2010 Conference Paper

t-logistic regression

Nan Ding
S. V. N. Vishwanathan

We extend logistic regression by using t-exponential families which were introduced recently in statistical physics. This gives rise to a regularized risk minimization problem with a non-convex loss function. An efficient block coordinate descent optimization scheme can be derived for estimating the parameters. Because of the nature of the loss function, our algorithm is tolerant to label noise. Furthermore, unlike other algorithms which employ non-convex loss functions, our algorithm is fairly robust to the choice of initial values. We verify both these observations empirically on a number of synthetic and real datasets.

PDF Details