Ziyu Chen

ICRA Conference 2025 Conference Paper

FGO-SLAM: Enhancing Gaussian SLAM with Globally Consistent Opacity Radiance Field

• Fan Zhu
• Yifan Zhao
• Ziyu Chen
• Biao Yu
• Hui Zhu 0010

Visual SLAM has regained attention due to its ability to provide perceptual capabilities and simulation test data for Embodied AI. However, traditional SLAM methods struggle to meet the demands of high-quality scene reconstruction, and Gaussian SLAM systems, despite their rapid rendering and high-quality mapping capabilities, lack effective pose optimization methods and face challenges in geometric reconstruction. To address these issues, we introduce FGO-SLAM, a Gaussian SLAM system that employs an opacity radiance field as the scene representation to enhance geometric mapping performance. After initial pose estimation, we apply global adjustment to optimize camera poses and sparse point cloud, ensuring robust tracking of our approach. Additionally, we maintain a globally consistent opacity radiance field based on 3D Gaussians and introduce depth distortion and normal consistency terms to refine the scene representation. Further-more, after constructing tetrahedral grids, we identify level sets to directly extract surfaces from 3D Gaussians. Results across various real-world and large-scale synthetic datasets demonstrate that our method achieves state-of-the-art tracking accuracy and mapping performance.

AAMAS Conference 2025 Conference Paper

Modeling the Collaborative Edge Data Caching Problem via a Dynamic DCOP

• Ziyang Song
• Ziyu Chen
• Jinhui Huang
• Cheng Zhang
• Jingyuan He

The Collaborative Edge Data Caching (CEDC) problem poses a significant challenge in Mobile Edge Computing (MEC). It’s a research focus to address the problem from the service providers’ perspective that requires the optimal caching strategy for service providers to maximize their caching revenue, subject to capacity and latency constraints. However, current research primarily focuses on centralized methods, neglecting the distributed and dynamic nature of CEDC. Accordingly, we first propose to use a Dynamic Distributed Constraint Optimization Problem (D-DCOP) to model the problem in a distributed manner, where capacity, latency constraints and caching revenue are dynamically mapped into local hard constraints and constraint utilities between edge servers according to changes in user requests. The proposed model enables each edge server to make its caching strategy through information exchange with neighboring edge servers. We further present a local search framework for CEDC to handle local hard constraints in the model and apply it to two classic local search algorithms, DSA and MGM, along with specific modifications to avoid repetitive computation. We empirically confirm the superiority of our distributed model and algorithms over state-of-the-art centralized solvers for CEDC.

ICLR Conference 2025 Conference Paper

OmniRe: Omni Urban Scene Reconstruction

• Ziyu Chen
• Jiawei Yang 0002
• Jiahui Huang
• Riccardo de Lutio
• Janick Martinez Esturo
• Boris Ivanovic
• Or Litany
• Zan Gojcic

We introduce OmniRe, a comprehensive system for efficiently creating high-fidelity digital twins of dynamic real-world scenes from on-device logs. Recent methods using neural fields or Gaussian Splatting primarily focus on vehicles, hindering a holistic framework for all dynamic foregrounds demanded by downstream applications, e.g., the simulation of human behavior. OmniRe extends beyond vehicle modeling to enable accurate, full-length reconstruction of diverse dynamic objects in urban scenes. Our approach builds scene graphs on 3DGS and constructs multiple Gaussian representations in canonical spaces that model various dynamic actors, including vehicles, pedestrians, cyclists, and others. OmniRe allows holistically reconstructing any dynamic object in the scene, enabling advanced simulations (~60 Hz) that include human-participated scenarios, such as pedestrian behavior simulation and human-vehicle interaction. This comprehensive simulation capability is unmatched by existing methods. Extensive evaluations on the Waymo dataset show that our approach outperforms prior state-of-the-art methods quantitatively and qualitatively by a large margin. We further extend our results to 5 additional popular driving datasets to demonstrate its generalizability on common urban scenes. Code and results are available at [omnire](https://ziyc.github.io/omnire/).

JAAMAS Journal 2024 Journal Article

Toward fast belief propagation for distributed constraint optimization problems via heuristic search

• Junsong Gao
• Ziyu Chen
• Qiang Li

Abstract Belief propagation (BP) approaches, such as Max-sum and its variants, are important methods to solve large-scale Distributed Constraint Optimization Problems. However, these algorithms face a huge challenge since their computational complexity scales exponentially with the arity of each constraint function. Current accelerating techniques for BP use sorting or branch-and-bound (BnB) strategy to reduce the search space. However, the existing BnB-based methods are mainly designed for specific problems, which limits their applicability. On the other hand, though several generic sorting-based methods have been proposed, they require significantly high preprocessing as well as memory overhead, which prohibits their adoption in some realistic scenarios. In this paper, we aim to propose a series of generic and memory-efficient heuristic search techniques to accelerate belief propagation. Specifically, by leveraging dynamic programming, we efficiently build function estimations for every partial assignment scoped in a constraint function in the preprocessing phase. Then, by using these estimations to build upper bounds and employing a branch-and-bound in a depth-first fashion to reduce the search space, we propose our first method called FDSP. Next, we enhance FDSP by adapting a concurrent-search strategy and leveraging the upper bounds as guiding information and propose its first heuristic variant framework called CONC-FDSP. Finally, by choosing to expand the partial assignment with the highest upper bound in each step of exploration, we propose the second heuristic variant of FDSP, called BFS-FDSP. We prove the correctness of our methods theoretically, and our empirical evaluations indicate their superiority for accelerating Max-sum in terms of both time and memory, compared with the state-of-the-art.

ICLR Conference 2023 Conference Paper

Mosaic Representation Learning for Self-supervised Visual Pre-training

• Zhaoqing Wang
• Ziyu Chen
• Yaqian Li
• Yandong Guo
• Jun Yu 0001
• Mingming Gong
• Tongliang Liu

Self-supervised learning has achieved significant success in learning visual representations without the need for manual annotation. To obtain generalizable representations, a meticulously designed data augmentation strategy is one of the most crucial parts. Recently, multi-crop strategies utilizing a set of small crops as positive samples have been shown to learn spatially structured features. However, it overlooks the diverse contextual backgrounds, which reduces the variance of the input views and degenerates the performance. To address this problem, we propose a mosaic representation learning framework (MosRep), consisting of a new data augmentation strategy that enriches the backgrounds of each small crop and improves the quality of visual representations. Specifically, we randomly sample numbers of small crops from different input images and compose them into a mosaic view, which is equivalent to introducing different background information for each small crop. Additionally, we further jitter the mosaic view to prevent memorizing the spatial locations of each crop. Along with optimization, our MosRep gradually extracts more discriminative features. Extensive experimental results demonstrate that our method improves the performance far greater than the multi-crop strategy on a series of downstream tasks, e.g., +7.4% and +4.9% than the multi-crop strategy on ImageNet-1K with 1% label and 10% label, respectively. Code is available at https://github.com/DerrickWang005/MosRep.git.

NeurIPS Conference 2023 Conference Paper

On the Implicit Bias of Linear Equivariant Steerable Networks

• Ziyu Chen
• Wei Zhu

We study the implicit bias of gradient flow on linear equivariant steerable networks in group-invariant binary classification. Our findings reveal that the parameterized predictor converges in direction to the unique group-invariant classifier with a maximum margin defined by the input group action. Under a unitary assumption on the input representation, we establish the equivalence between steerable networks and data augmentation. Furthermore, we demonstrate the improved margin and generalization bound of steerable networks over their non-invariant counterparts.

ICML Conference 2023 Conference Paper

Sample Complexity of Probability Divergences under Group Symmetry

• Ziyu Chen
• Markos A. Katsoulakis
• Luc Rey-Bellet
• Wei Zhu 0007

We rigorously quantify the improvement in the sample complexity of variational divergence estimations for group-invariant distributions. In the cases of the Wasserstein-1 metric and the Lipschitz-regularized $\alpha$-divergences, the reduction of sample complexity is proportional to an ambient-dimension-dependent power of the group size. For the maximum mean discrepancy (MMD), the improvement of sample complexity is more nuanced, as it depends on not only the group size but also the choice of kernel. Numerical simulations verify our theories.

AAMAS Conference 2022 Conference Paper

Beyond Uninformed Search: Improving Branch-and-bound Based Acceleration Algorithms for Belief Propagation via Heuristic Strategies

• Junsong Gao
• Ziyu Chen
• Dingding Chen
• Wenxin Zhang

Belief propagation algorithms including Max-sum and its variants are important methods for solving DCOPs. However, they may face a tough challenge when handling n-ary constraints since the computational overheads grow exponentially with the number of variables that a utility function holds. In this paper, we update the state-of-the-art technique called Function Decomposing and State Pruning (FDSP) which can significantly reduce such an expenditure, by introducing two heuristic techniques. By introducing a round-robin mechanism to control the order of exploration, we propose Concurrent-search-based FDSP (CONC-FDSP). Besides, we propose Best-first-search-based FDSP (BFS-FDSP) by using the 𝐴∗ search to find the optimal path to the solution. Finally, we demonstrate their efficiency in solving the benchmarks compared with the state-of-the-art.

TCS Journal 2022 Journal Article

Constructions of 2-resilient rotation symmetric Boolean functions through symbol transformations of cyclic Hadamard matrix

• Jiao Du
• Ziyu Chen
• Shaojing Fu
• Longjiang Qu
• Chao Li

In this paper, the properties of symbol transformations of cyclic Hadamard matrices are studied. An infinite class of ( n − 1 ) -variable 2-resilient rotation symmetric Boolean functions are constructed, and the nonlinearity of the constructed functions is 2 n ( n − 1 ). The crucial technique of this method is to determine a subset T ⊆ F 2 n − 1 satisfying a correspondent condition. This is a new construction of 2-resilient rotation symmetric Boolean functions via switching the supports of ( n − 1 ) -variable rotation symmetric Boolean functions of degree one, i. e. , f 0 n − 1 ( x 1, x 2, ⋯, x n − 1 ) = ⊕ i = 1 n − 1 x i, where n = 4 t.

ICLR Conference 2022 Conference Paper

Representation-Agnostic Shape Fields

• Xiaoyang Huang
• Jiancheng Yang
• Yanjun Wang
• Ziyu Chen
• Linguo Li
• Teng Li 0001
• Bingbing Ni
• Wenjun Zhang 0001

3D shape analysis has been widely explored in the era of deep learning. Numerous models have been developed for various 3D data representation formats, e.g., MeshCNN for meshes, PointNet for point clouds and VoxNet for voxels. In this study, we present Representation-Agnostic Shape Fields (RASF), a generalizable and computation-efficient shape embedding module for 3D deep learning. RASF is implemented with a learnable 3D grid with multiple channels to store local geometry. Based on RASF, shape embeddings for various 3D shape representations (point clouds, meshes and voxels) are retrieved by coordinate indexing. While there are multiple ways to optimize the learnable parameters of RASF, we provide two effective schemes among all in this paper for RASF pre-training: shape reconstruction and normal estimation. Once trained, RASF becomes a plug-and-play performance booster with negligible cost. Extensive experiments on diverse 3D representation formats, networks and applications, validate the universal effectiveness of the proposed RASF. Code and pre-trained models are publicly available\footnote{\url{https://github.com/seanywang0408/RASF}}.

JAAMAS Journal 2020 Journal Article

A genetic algorithm based framework for local search algorithms for distributed constraint optimization problems

• Ziyu Chen
• Lizhen Liu
• Zhepeng Yu

Abstract Local search algorithms are widely applied in solving large-scale Distributed constraint optimization problems (DCOPs) where each agent holds a value assignment to its variable and iteratively makes a decision on whether to replace its assignment according to its neighbor states. However, the value assignments of their neighbors confine their search to a small space so that agents in local search algorithms easily fall into local optima. Fortunately, Genetic Algorithms (GAs) can direct a search process to a more promising space and help the search process to break up the confine of local states. Accordingly, we propose a GA-based framework (LSGA) to enhance local search algorithms, where a series of genetic operators are redesigned for agents in distributed scenario to accommodate DCOPs. First, a fitness function is designed to evaluate the assignments for each agent, considering the balance of local benefits and global benefits. Then, a new method is provided to decide crossover positions in terms of agent-communication and topological structure of DCOPs. Besides, a self-adaptive crossover probability and a self-adaptive mutation probability are proposed to control the uses of crossover operator and mutation operator, respectively. And more importantly, the LSGA framework can be easily applied in any local search algorithm. The experimental results demonstrate the superiority of the use of LSGA in the typical search algorithms over state-of-the-art incomplete algorithms.

AAAI Conference 2020 Conference Paper

HS-CAI: A Hybrid DCOP Algorithm via Combining Search with Context-Based Inference

• Dingding Chen
• Yanchen Deng
• Ziyu Chen
• Wenxing Zhang
• Zhongshi He

Search and inference are two main strategies for optimally solving Distributed Constraint Optimization Problems (DCOPs). Recently, several algorithms were proposed to combine their advantages. Unfortunately, such algorithms only use an approximated inference as a one-shot preprocessing phase to construct the initial lower bounds which lead to inefficient pruning under the limited memory budget. On the other hand, iterative inference algorithms (e. g. , MB-DPOP) perform a context-based complete inference for all possible contexts but suffer from tremendous traffic overheads. In this paper, (i) hybridizing search with context-based inference, we propose a complete algorithm for DCOPs, named HS- CAI where the inference utilizes the contexts derived from the search process to establish tight lower bounds while the search uses such bounds for efficient pruning and thereby reduces contexts for the inference. Furthermore, (ii) we introduce a context evaluation mechanism to select the context patterns for the inference to further reduce the overheads incurred by iterative inferences. Finally, (iii) we prove the correctness of our algorithm and the experimental results demonstrate its superiority over the state-of-the-art.

AAAI Conference 2019 Conference Paper

A Generic Approach to Accelerating Belief Propagation Based Incomplete Algorithms for DCOPs via a Branch-and-Bound Technique

• Ziyu Chen
• Xingqiong Jiang
• Yanchen Deng
• Dingding Chen
• Zhongshi He

Belief propagation approaches, such as Max-Sum and its variants, are important methods to solve large-scale Distributed Constraint Optimization Problems (DCOPs). However, for problems with n-ary constraints, these algorithms face a huge challenge since their computational complexity scales exponentially with the number of variables a function holds. In this paper, we present a generic and easy-touse method based on a branch-and-bound technique to solve the issue, called Function Decomposing and State Pruning (FDSP). We theoretically prove that FDSP can provide monotonically non-increasing upper bounds and speed up belief propagation based incomplete DCOP algorithms without an effect on solution quality. Also, our empirically evaluation indicates that FDSP can reduce 97% of the search space at least and effectively accelerate Max-Sum, compared with the state-of-the-art.

IJCAI Conference 2019 Conference Paper

AsymDPOP: Complete Inference for Asymmetric Distributed Constraint Optimization Problems

• Yanchen Deng
• Ziyu Chen
• Dingding Chen
• Wenxin Zhang
• Xingqiong Jiang

Asymmetric distributed constraint optimization problems (ADCOPs) are an emerging model for coordinating agents with personal preferences. However, the existing inference-based complete algorithms which use local eliminations cannot be applied to ADCOPs, as the parent agents are required to transfer their private functions to their children. Rather than disclosing private functions explicitly to facilitate local eliminations, we solve the problem by enforcing delayed eliminations and propose AsymDPOP, the first inference-based complete algorithm for ADCOPs. To solve the severe scalability problems incurred by delayed eliminations, we propose to reduce the memory consumption by propagating a set of smaller utility tables instead of a joint utility table, and to reduce the computation efforts by sequential optimizations instead of joint optimizations. The empirical evaluation indicates that AsymDPOP significantly outperforms the state-of-the-art, as well as the vanilla DPOP with PEAV formulation.

AAMAS Conference 2019 Conference Paper

PT-ISABB: A Hybrid Tree-based Complete Algorithm to Solve Asymmetric Distributed Constraint Optimization Problems

• Yanchen Deng
• Ziyu Chen
• Dingding Chen
• Xingqiong Jiang
• Qiang Li

Asymmetric Distributed Constraint Optimization Problems (AD- COPs) have emerged as an important formalism in multi-agent community due to their ability to capture personal preferences. However, the existing search-based complete algorithms for AD- COPs can only use local knowledge to compute lower bounds, which leads to inefficient pruning and prohibits them from solving large scale problems. On the other hand, inference-based complete algorithms (e. g. , DPOP) for Distributed Constraint Optimization Problems (DCOPs) require only a linear number of messages, but they cannot be directly applied into ADCOPs due to a privacy concern. Therefore, in the paper, we consider the possibility of combining inference and search to effectively solve ADCOPs at an acceptable loss of privacy. Specifically, we propose a hybrid complete algorithm called PT-ISABB which uses a tailored inference algorithm to provide tight lower bounds and a tree-based complete search algorithm to exhaust the search space. We prove the correctness of our algorithm and the experimental results demonstrate its superiority over other state-of-the-art complete algorithms.

JAAMAS Journal 2018 Journal Article

A class of iterative refined Max-sum algorithms via non-consecutive value propagation strategies

• Ziyu Chen
• Yanchen Deng
• Zhongshi He

Abstract As an important technique to solve distributed constraint optimization problems, Max-sum has drawn a lot of attention and successfully been deployed in real applications. Unfortunately, Max-sum fails to converge in cyclic problems and usually traverses states with low quality. Max-sum_AD and Max-sum_ADVP were proposed to guarantee the single phase convergence and the cross phase convergence respectively, and greatly improve the solution quality of Max-sum. However, the solution quality is closely related to the timing for starting value propagation in Max-sum_ADVP. In other words, low-quality initial assignments will lead to a poor result. In this paper, we prove that value propagation could restrict the exploration ability brought by Max-sum and eventually makes Max-sum_ADVP equivalent to a sequential greedy local search algorithm. For getting a balance between exploration and exploitation, several non-consecutive value propagation strategies are proposed to relax the restriction caused by value propagation: single-side value propagation which executes value propagation and Max-sum_AD in an interleaved way, probabilistic value propagation which performs value propagation stochastically and hybrid belief/value propagation where agents perform Max-sum_AD and value propagation in one round. We illustrate that agents in our algorithms can make decisions beyond local functions. Our empirical evaluations demonstrate the superiority of our methods over Max-sum and its variants. It also can be found that our methods are independent of the value propagation timing which is a major concern in Max-sum_ADVP.

AAAI Conference 2018 Conference Paper

An Ant-Based Algorithm to Solve Distributed Constraint Optimization Problems

• Ziyu Chen
• Tengfei Wu
• Yanchen Deng
• Cheng Zhang

As an important population-based algorithm, ant colony optimization (ACO) has been successfully applied into various combinatorial optimization problems. However, much existing work in ACO focuses on solving centralized problems. In this paper, we present a novel algorithm that takes the power of ants to solve Distributed Constraint Optimization Problems (DCOPs), called ACO DCOP. In ACO DCOP, a new mechanism that captures local benefits is proposed to compute heuristic factors and a new method that considers the cost structure of DCOPs is proposed to compute pheromone deltas appropriately. Moreover, pipelining technique is introduced to make full use of the computational capacity and improve the efficiency. In our theoretical analysis, we prove that ACO DCOP is an anytime algorithm. Our empirical evaluation indicates that ACO DCOP is able to find solutions of equal or significantly higher quality than state-of-the-art DCOP algorithms.

AAMAS Conference 2017 Conference Paper

A Partial Decision Scheme for Local Search Algorithms for Distributed Constraint Optimization Problems

• Zhepeng Yu
• Ziyu Chen
• Jingyuan He
• Yancheng Deng

AAMAS Conference 2017 Conference Paper

An Iterative Refined Max-sum_AD Algorithm via Single-side Value Propagation and Local Search

• Ziyu Chen
• Yanchen Deng
• Tengfei Wu

Max-sum ADVP is a message-passing algorithm for solving Distributed Constraint Optimization Problems (DCOPs) able to obtain a convergent solution in a cyclic factor graph. Nevertheless, the solution quality is closely related to the timing for starting value propagation in Max-sum ADVP. In other words, low-quality initial assignments will lead to a poor result. In this paper, we illustrate that value propagation can eliminate the inconsistent contexts in Max-sum AD and break ties among utilities, but it also restricts the exploration brought by Max-sum. For balancing between the exploration and the accuracy, we propose a new iterative refined Max-sum AD algorithm with single-side value propagation, called Max-sum ADSSVP. It performs two phases in every two convergences, one phase which enables the exploration to find high-quality initial assignments and the other phase which enables value propagation to guarantee solution quality. Max-sum ADSSVP tackles the timing selection problem by iteratively refining initial assignments in every exploration phase. Besides, local search is introduced after the value propagation phase to speed up the convergence process. Our empirical evaluation indicates that our methods are independent of initial assignments and less likely to get stuck in local optima.