Author name cluster

Ziyu Wan

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.

13 papers

2 author rows

IROS Conference 2025 Conference Paper

DnD Filter: Differentiable State Estimation for Dynamic Systems using Diffusion Models

Ziyu Wan
Lin Zhao

This paper proposes the DnD Filter, a differentiable filter that utilizes diffusion models for state estimation of dynamic systems. Unlike conventional differentiable filters, which often impose restrictive assumptions on process noise (e. g. , Gaussianity), DnD Filter enables a nonlinear state update without such constraints by conditioning a diffusion model on both the predicted state and observational data, capitalizing on its ability to approximate complex distributions. We validate its effectiveness on both a simulated task and a real-world visual odometry task, where DnD Filter consistently outperforms existing baselines. Specifically, it achieves a 25% improvement in estimation accuracy on the visual odometry task compared to state-of-the-art differentiable filters, and even surpasses differentiable smoothers that utilize future measurements. To the best of our knowledge, DnD Filter represents the first successful attempt to leverage diffusion models for state estimation, offering a flexible and powerful framework for nonlinear estimation under noisy measurements. The code is available at https://github.com/ZiyuNUS/DnDFilter.

Details

ICLR Conference 2025 Conference Paper

EG4D: Explicit Generation of 4D Object without Score Distillation

Qi Sun
Zhiyang Guo
Ziyu Wan
Jing Nathan Yan
Shengming Yin
Wengang Zhou 0001
Jing Liao 0001
Houqiang Li

In recent years, the increasing demand for dynamic 3D assets in design and gaming applications has given rise to powerful generative pipelines capable of synthesizing high-quality 4D objects. Previous methods generally rely on score distillation sampling (SDS) algorithm to infer the unseen views and motion of 4D objects, thus leading to unsatisfactory results with defects like over-saturation and Janus problem. Therefore, inspired by recent progress of video diffusion models, we propose to optimize a 4D representation by explicitly generating multi-view videos from one input image. However, it is far from trivial to handle practical challenges faced by such a pipeline, including dramatic temporal inconsistency, inter-frame geometry and texture diversity, and semantic defects brought by video generation results. To address these issues, we propose EG4D, a novel multi-stage framework that generates high-quality and consistent 4D assets without score distillation. Specifically, collaborative techniques and solutions are developed, including an attention injection strategy to synthesize temporal-consistent multi-view videos, a robust and efficient dynamic reconstruction method based on Gaussian Splatting, and a refinement stage with diffusion prior for semantic restoration. The qualitative comparisons and quantitative results demonstrate that our framework outperforms the baselines in generation quality by a considerable margin.

Details

ICML Conference 2025 Conference Paper

Epsilon-VAE: Denoising as Visual Decoding

Long Zhao 0003
Sanghyun Woo
Ziyu Wan
Yandong Li
Han Zhang 0010
Boqing Gong
Hartwig Adam
Xuhui Jia

In generative modeling, tokenization simplifies complex data into compact, structured representations, creating a more efficient, learnable space. For high-dimensional visual data, it reduces redundancy and emphasizes key features for high-quality generation. Current visual tokenization methods rely on a traditional autoencoder framework, where the encoder compresses data into latent representations, and the decoder reconstructs the original input. In this work, we offer a new perspective by proposing denoising as decoding, shifting from single-step reconstruction to iterative refinement. Specifically, we replace the decoder with a diffusion process that iteratively refines noise to recover the original image, guided by the latents provided by the encoder. We evaluate our approach by assessing both reconstruction (rFID) and generation quality (FID), comparing it to state-of-the-art autoencoding approaches. By adopting iterative reconstruction through diffusion, our autoencoder, namely Epsilon-VAE, achieves high reconstruction quality, which in turn enhances downstream generation quality by 22% at the same compression rates or provides 2. 3x inference speedup through increasing compression rates. We hope this work offers new insights into integrating iterative generation and autoencoding for improved compression and generation.

Details

NeurIPS Conference 2025 Conference Paper

ReMA: Learning to Meta-Think for LLMs with Multi-agent Reinforcement Learning

Ziyu Wan
Yunxiang Li
Xiaoyu Wen
Yan Song
Hanjing Wang
Linyi Yang
Mark Schmidt
Jun Wang

Recent research on Reasoning of Large Language Models (LLMs) has sought to further enhance their performance by integrating meta-thinking—enabling models to monitor, evaluate, and control their reasoning processes for more adaptive and effective problem-solving. However, current single-agent work lacks a specialized design for acquiring meta-thinking, resulting in low efficacy. To address this challenge, we introduce Reinforced Meta-thinking Agents (ReMA), a novel framework that leverages Multi-Agent Reinforcement Learning (MARL) to elicit meta-thinking behaviors, encouraging LLMs to think about thinking. ReMA decouples the reasoning process into two hierarchical agents: a high-level meta-thinking agent responsible for generating strategic oversight and plans, and a low-level reasoning agent for detailed executions. Through iterative reinforcement learning with aligned objectives, these agents explore and learn collaboration, leading to improved generalization and robustness. Empirical results from single-turn experiments demonstrate that ReMA outperforms single-agent RL baselines on complex reasoning tasks, including competitive-level mathematical benchmarks and LLM-as-a-Judge benchmarks. Additionally, we further extend ReMA to multi-turn interaction settings, leveraging turn-level ratio and parameter sharing to improve efficiency. Comprehensive ablation studies further illustrate the evolving dynamics of each distinct agent, providing valuable insights into how the meta-thinking reasoning process enhances the reasoning capabilities of LLMs.

PDF Details

NeurIPS Conference 2025 Conference Paper

ThinkBench: Dynamic Out-of-Distribution Evaluation for Robust LLM Reasoning

Shulin Huang
Linyi Yang
Yan Song
Shawn Chen
Leyang Cui
Ziyu Wan
Qingcheng Zeng
Ying Wen

Evaluating large language models (LLMs) poses significant challenges, particularly due to issues of data contamination and the leakage of correct answers. To address these challenges, we introduce ThinkBench, a novel evaluation framework designed to robustly evaluate the reasoning capability of LLMs. ThinkBench proposes a dynamic data generation method for constructing out-of-distribution (OOD) datasets and offers an OOD dataset that contains 2, 912 samples drawn from reasoning tasks. ThinkBench unifies the evaluation of reasoning models and non-reasoning models. We evaluate 16 LLMs and 4 PRMs under identical experimental conditions and show that most of the LLMs' performance are far from robust and they face a certain level of data leakage. By dynamically generating OOD datasets, ThinkBench effectively provides a reliable evaluation of LLMs and reduces data contamination impact. Our data and codes are available at https: //github. com/huangshulin123/ThinkBench.

PDF Details

ICML Conference 2024 Conference Paper

AlphaZero-Like Tree-Search can Guide Large Language Model Decoding and Training

Ziyu Wan
Xidong Feng
Muning Wen
Stephen Marcus McAleer
Ying Wen 0001
Weinan Zhang 0001
Jun Wang 0012

Recent works like Tree-of-Thought (ToT) and Reasoning via Planning (RAP) aim to augment the multi-step reasoning capabilities of LLMs by using tree-search algorithms. These methods rely on prompting a pre-trained model to serve as a value function and focus on problems with low search depth. As a result, these methods cannot benefit from in-domain training and only rely on pretraining process — they will not work in domains where the pre-trained LLM does not have enough knowledge to serve as an effective value function or in domains that require long-horizon planning. To address these limitations, we present an AlphaZero-like tree-search learning framework for LLMs (termed TS-LLM), systematically illustrating how tree-search with a learned value function can guide LLM decoding. TS-LLM distinguishes itself in two key ways. (1) Leveraging a learned value function and AlphaZero-like algorithms, our approach can be generally adaptable to a wide range of tasks, language models of any size, and tasks of varying search depths. (2) Our approach can guide LLMs during both inference and training, iteratively improving the LLMs. Empirical results across reasoning, planning, alignment, and decision-making tasks show that TS-LLM outperforms existing approaches and can handle trees with a depth of 64.

Details

NeurIPS Conference 2024 Conference Paper

Reinforcing LLM Agents via Policy Optimization with Action Decomposition

Muning Wen
Ziyu Wan
Jun Wang
Weinan Zhang
Ying Wen

Language models as intelligent agents push the boundaries of sequential decision-making agents but struggle with limited knowledge of environmental dynamics and exponentially huge action space. Recent efforts like GLAM and TWOSOME manually constrain the action space to a restricted subset and employ reinforcement learning to align agents' knowledge with specific environments. However, they overlook fine-grained credit assignments for intra-action tokens, which is essential for efficient language agent optimization, and rely on human's prior knowledge to restrict action space. This paper proposes decomposing language agent optimization from the action level to the token level, offering finer supervision for each intra-action token and manageable optimization complexity in environments with unrestricted action spaces. Beginning with the simplification of flattening all actions, we theoretically explore the discrepancies between action-level optimization and this naive token-level optimization. We then derive the Bellman backup with Action Decomposition (BAD) to integrate credit assignments for both intra-action and inter-action tokens, effectively eliminating the discrepancies. Implementing BAD within the PPO algorithm, we introduce Policy Optimization with Action Decomposition (POAD). POAD benefits from a finer-grained credit assignment process and lower optimization complexity, leading to enhanced learning efficiency and generalization abilities in aligning language agents with interactive environments. We validate POAD across diverse testbeds, with results affirming the advantages of our approach and the correctness of our theoretical analysis. The source code can be accessed directly with this link: https: //github. com/morning9393/ADRL.

PDF Details DOI

JMLR Journal 2023 Journal Article

MALib: A Parallel Framework for Population-based Multi-agent Reinforcement Learning

Ming Zhou
Ziyu Wan
Hanjing Wang
Muning Wen
Runzhe Wu
Ying Wen
Yaodong Yang
Yong Yu

Population-based multi-agent reinforcement learning (PB-MARL) encompasses a range of methods that merge dynamic population selection with multi-agent reinforcement learning algorithms (MARL). While PB-MARL has demonstrated notable achievements in complex multi-agent tasks, its sequential execution is plagued by low computational efficiency due to the diversity in computing patterns and policy combinations. We propose a solution involving a stateless central task dispatcher and stateful workers to handle PB-MARL's subroutines, thereby capitalizing on parallelism across various components for efficient problem-solving. In line with this approach, we introduce MALib, a parallel framework that incorporates a task control model, independent data servers, and an abstraction of MARL training paradigms. The framework has undergone extensive testing and is available under the MIT license (https://github.com/sjtu-marl/malib) [abs] [ pdf ][ bib ] [ code ] &copy JMLR 2023. ( edit, beta )

PDF Details

ICLR Conference 2023 Conference Paper

Order Matters: Agent-by-agent Policy Optimization

Xihuai Wang
Zheng Tian 0002
Ziyu Wan
Ying Wen 0001
Jun Wang 0012
Weinan Zhang 0001

While multi-agent trust region algorithms have achieved great success empirically in solving coordination tasks, most of them, however, suffer from a non-stationarity problem since agents update their policies simultaneously. In contrast, a sequential scheme that updates policies agent-by-agent provides another perspective and shows strong performance. However, sample inefficiency and lack of monotonic improvement guarantees for each agent are still the two significant challenges for the sequential scheme. In this paper, we propose the \textbf{A}gent-by-\textbf{a}gent \textbf{P}olicy \textbf{O}ptimization (A2PO) algorithm to improve the sample efficiency and retain the guarantees of monotonic improvement for each agent during training. We justify the tightness of the monotonic improvement bound compared with other trust region algorithms. From the perspective of sequentially updating agents, we further consider the effect of agent updating order and extend the theory of non-stationarity into the sequential update scheme. To evaluate A2PO, we conduct a comprehensive empirical study on four benchmarks: StarCraftII, Multi-agent MuJoCo, Multi-agent Particle Environment, and Google Research Football full game scenarios. A2PO consistently outperforms strong baselines.

Details

AAAI Conference 2021 Short Paper

Global Fusion Attention for Vision and Language Understanding (Student Abstract)

Zixin Guo
Chen Liang
Ziyu Wan
Yang Bai

We extend the popular transformer architecture to a multimodal model, processing both visual and textual inputs. We propose a new attention mechanism on Transformer-based architecture for the joint vision and language understanding tasks. Our model fuses multi-level comprehension between images and texts in a weighted manner, which could better curve the internal relationships. Experiments on benchmark VQA dataset CLEVR demonstrate the effectiveness of the proposed attention mechanism. We also observe the improvements in sample efficiency of reinforcement learning through the experiments on grounded language understanding tasks of BabyAI platform.

PDF Details

NeurIPS Conference 2021 Conference Paper

Neural Auto-Curricula in Two-Player Zero-Sum Games

Xidong Feng
Oliver Slumbers
Ziyu Wan
Bo Liu
Stephen McAleer
Ying Wen
Jun Wang
Yaodong Yang

When solving two-player zero-sum games, multi-agent reinforcement learning (MARL) algorithms often create populations of agents where, at each iteration, a new agent is discovered as the best response to a mixture over the opponent population. Within such a process, the update rules of "who to compete with" (i. e. , the opponent mixture) and "how to beat them" (i. e. , finding best responses) are underpinned by manually developed game theoretical principles such as fictitious play and Double Oracle. In this paper, we introduce a novel framework—Neural Auto-Curricula (NAC)—that leverages meta-gradient descent to automate the discovery of the learning update rule without explicit human design. Specifically, we parameterise the opponent selection module by neural networks and the best-response module by optimisation subroutines, and update their parameters solely via interaction with the game engine, where both players aim to minimise their exploitability. Surprisingly, even without human design, the discovered MARL algorithms achieve competitive or even better performance with the state-of-the-art population-based game solvers (e. g. , PSRO) on Games of Skill, differentiable Lotto, non-transitive Mixture Games, Iterated Matching Pennies, and Kuhn Poker. Additionally, we show that NAC is able to generalise from small games to large games, for example training on Kuhn Poker and outperforming PSRO on Leduc Poker. Our work inspires a promising future direction to discover general MARL algorithms solely from data.

PDF Details

AAAI Conference 2019 Short Paper

Transductive Zero-Shot Learning via Visual Center Adaptation

Ziyu Wan
Yan Li
Min Yang
Junge Zhang

In this paper, we propose a Visual Center Adaptation Method (VCAM) to address the domain shift problem in zero-shot learning. For the seen classes in the training data, VCAM builds an embedding space by learning the mapping from semantic space to some visual centers. While for unseen classes in the test data, the construction of embedding space is constrained by a symmetric Chamfer-distance term, aiming to adapt the distribution of the synthetic visual centers to that of the real cluster centers. Therefore the learned embedding space can generalize the unseen classes well. Experiments on two widely used datasets demonstrate that our model significantly outperforms state-of-the-art methods.

PDF Details

NeurIPS Conference 2019 Conference Paper

Transductive Zero-Shot Learning with Visual Structure Constraint

Ziyu Wan
DongDong Chen
Yan Li
Xingguang Yan
Junge Zhang
Yizhou Yu
Jing Liao

To recognize objects of the unseen classes, most existing Zero-Shot Learning (ZSL) methods first learn a compatible projection function between the common semantic space and the visual space based on the data of source seen classes, then directly apply it to the target unseen classes. However, in real scenarios, the data distribution between the source and target domain might not match well, thus causing the well-known domain shift problem. Based on the observation that visual features of test instances can be separated into different clusters, we propose a new visual structure constraint on class centers for transductive ZSL, to improve the generality of the projection function (\ie alleviate the above domain shift problem). Specifically, three different strategies (symmetric Chamfer-distance, Bipartite matching distance, and Wasserstein distance) are adopted to align the projected unseen semantic centers and visual cluster centers of test instances. We also propose a new training strategy to handle the real cases where many unrelated images exist in the test dataset, which is not considered in previous methods. Experiments on many widely used datasets demonstrate that the proposed visual structure constraint can bring substantial performance gain consistently and achieve state-of-the-art results.

PDF Details