Minhao Cheng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Minhao Cheng has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Hongyang Chen
Zhejiang Lab
4 shared publications - Cho-Jui Hsieh
Computer Science Department University of California Los Angeles CA USA
3 shared publications - Han Liu
Dalian University of Technology
2 shared publications - Tianyi Zhou
Mohamed bin Zayed University of Artificial Intelligence, Masdar City, Abu Dhabi, United Arab Emirates
2 shared publications - Andrew Bai
UCLA
1 shared publication - Bolin Ding
Alibaba Group, Bellevue, USA
1 shared publication
Latest Publications
- Open Access
- 20 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 1138-1149
Input Snapshots Fusion for Scalable Discrete-Time Dynamic Graph Neural Networks
- 20 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 1138-1149
Open AccessInput Snapshots Fusion for Scalable Discrete-Time Dynamic Graph Neural Networks
- Open Access
Exploring Expert Failures Improves LLM Agent Tuning
Large Language Models (LLMs) have shown tremendous potential as agents, excelling at tasks that require multiple rounds of reasoning and interactions. Rejection Sampling Fine-Tuning (RFT) has emerged as an effective method for finetuning LLMs as agents: it first imitates expert-generated successful trajectories and further improves agentic skills through iterative fine-tuning on successful, self-generated trajectories. However, since the expert (e.g., GPT-4) succeeds primarily on simpler subtasks and RFT inherently favors simpler scenarios, many complex subtasks remain unsolved and persistently out-of-distribution (OOD). Upon investigating these challenging subtasks, we discovered that previously failed expert trajectories can often provide valuable guidance, e.g., plans and key actions, that can significantly improve agent exploration efficiency and acquisition of critical skills. Motivated by these observations, we propose Exploring Expert Failures (EEF), which identifies beneficial actions from failed expert trajectories and integrates them into the training dataset. Potentially harmful actions are meticulously excluded to prevent contamination of the model learning process. By leveraging the beneficial actions in expert failures, EEF successfully solves some previously unsolvable subtasks and improves agent tuning performance. Remarkably, our approach achieved a 62\% win rate in WebShop, outperforming RFT (53. 6\%) and GPT-4 (35. 6\%), and to the best of our knowledge, setting a new state-of-the-art as the first method to surpass a score of 0.81 in WebShop and exceed 81 in SciWorld.
Open AccessExploring Expert Failures Improves LLM Agent Tuning
Large Language Models (LLMs) have shown tremendous potential as agents, excelling at tasks that require multiple rounds of reasoning and interactions. Rejection Sampling Fine-Tuning (RFT) has emerged as an effective method for finetuning LLMs as agents: it first imitates expert-generated successful trajectories and further improves agentic skills through iterative fine-tuning on successful, self-generated trajectories. However, since the expert (e.g., GPT-4) succeeds primarily on simpler subtasks and RFT inherently favors simpler scenarios, many complex subtasks remain unsolved and persistently out-of-distribution (OOD). Upon investigating these challenging subtasks, we discovered that previously failed expert trajectories can often provide valuable guidance, e.g., plans and key actions, that can significantly improve agent exploration efficiency and acquisition of critical skills. Motivated by these observations, we propose Exploring Expert Failures (EEF), which identifies beneficial actions from failed expert trajectories and integrates them into the training dataset. Potentially harmful actions are meticulously excluded to prevent contamination of the model learning process. By leveraging the beneficial actions in expert failures, EEF successfully solves some previously unsolvable subtasks and improves agent tuning performance. Remarkably, our approach achieved a 62\% win rate in WebShop, outperforming RFT (53. 6\%) and GPT-4 (35. 6\%), and to the best of our knowledge, setting a new state-of-the-art as the first method to surpass a score of 0.81 in WebShop and exceed 81 in SciWorld.
- 7 December 2024
- Published by Springer Nature
- p. 104-120
A Watermark-Conditioned Diffusion Model for IP Protection
- 7 December 2024
- Published by Springer Nature
- p. 104-120
A Watermark-Conditioned Diffusion Model for IP Protection
- 24 November 2024
- Published by Springer Nature
- p. 190-206
Understanding the Impact of Negative Prompts: When and How Do They Take Effect?
- 24 November 2024
- Published by Springer Nature
- p. 190-206
Understanding the Impact of Negative Prompts: When and How Do They Take Effect?
- Open Access
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 1785-1794
Exploring Robustness of GNN against Universal Injection Attack from a Worst-case Perspective
Recently, graph neural networks (GNNs) have demonstrated outstanding performance in fundamental tasks such as node classification and link prediction, as well as in specialized domains like recommendation systems, fraud detection, and drug discovery. However, their vulnerability to adversarial attacks raises concerns about their reliability in security-critical areas. To address this issue, researchers are exploring various defense methods, including specific attack countermeasures and certifiable robustness approaches. Nevertheless, these strategies are often effective only against limited attack scenarios, and prevailing certification methods prove inadequate when confronted with injection attacks. In this paper, we propose a method named CERT_UIA to enhance the robustness of GNN models against worst-case attacks, specifically targeting the scenario of Universal node Injection Attacks (UIA), thereby filling a gap in the existing literature on certified robustness in this context. Our approach involves a two-stage attack process that replaces the transformations of the topology and feature spaces with equivalent unified feature transformations, unifying the optimization of worst-case perturbations into a single feature space. Furthermore, we empirically evaluate our method on several benchmark datasets and compare it with existing certified methods.
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 1785-1794
Open AccessExploring Robustness of GNN against Universal Injection Attack from a Worst-case Perspective
Recently, graph neural networks (GNNs) have demonstrated outstanding performance in fundamental tasks such as node classification and link prediction, as well as in specialized domains like recommendation systems, fraud detection, and drug discovery. However, their vulnerability to adversarial attacks raises concerns about their reliability in security-critical areas. To address this issue, researchers are exploring various defense methods, including specific attack countermeasures and certifiable robustness approaches. Nevertheless, these strategies are often effective only against limited attack scenarios, and prevailing certification methods prove inadequate when confronted with injection attacks. In this paper, we propose a method named CERT_UIA to enhance the robustness of GNN models against worst-case attacks, specifically targeting the scenario of Universal node Injection Attacks (UIA), thereby filling a gap in the existing literature on certified robustness in this context. Our approach involves a two-stage attack process that replaces the transformations of the topology and feature spaces with equivalent unified feature transformations, unifying the optimization of worst-case perturbations into a single feature space. Furthermore, we empirically evaluate our method on several benchmark datasets and compare it with existing certified methods.
- Open Access
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 3057-3068
Unsupervised Heterogeneous Graph Rewriting Attack via Node Clustering
Self-supervised learning (SSL) has become one of the most popular learning paradigms and has achieved remarkable success in the graph field. Recently, a series of pre-training studies on heterogeneous graphs (HGs) using SSL have been proposed considering the heterogeneity of real-world graph data. However, verification of the robustness of heterogeneous graph pre-training is still a research gap. Most existing researches focus on supervised attacks on graphs, which are limited to a specific scenario and will not work when labels are not available. In this paper, we propose a novel unsupervised heterogeneous graph rewriting attack via node clustering (HGAC) that can effectively attack HG pre-training models without using labels. Specifically, a heterogeneous edge rewriting strategy is designed to ensure the rationality and concealment of the attacks. Then, a tailored heterogeneous graph contrastive learning (HGCL) is used as a surrogate model. Moreover, we leverage node clustering results of the clean HGs as the pseudo-labels to guide the optimization of structural attacks. Extensive experiments exhibit powerful attack performances of our HGAC on various downstream tasks (i.e., node classification, node clustering, metapath prediction, and visualization) under poisoning attack and evasion attack.
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 3057-3068
Open AccessUnsupervised Heterogeneous Graph Rewriting Attack via Node Clustering
Self-supervised learning (SSL) has become one of the most popular learning paradigms and has achieved remarkable success in the graph field. Recently, a series of pre-training studies on heterogeneous graphs (HGs) using SSL have been proposed considering the heterogeneity of real-world graph data. However, verification of the robustness of heterogeneous graph pre-training is still a research gap. Most existing researches focus on supervised attacks on graphs, which are limited to a specific scenario and will not work when labels are not available. In this paper, we propose a novel unsupervised heterogeneous graph rewriting attack via node clustering (HGAC) that can effectively attack HG pre-training models without using labels. Specifically, a heterogeneous edge rewriting strategy is designed to ensure the rationality and concealment of the attacks. Then, a tailored heterogeneous graph contrastive learning (HGCL) is used as a surrogate model. Moreover, we leverage node clustering results of the clean HGs as the pseudo-labels to guide the optimization of structural attacks. Extensive experiments exhibit powerful attack performances of our HGAC on various downstream tasks (i.e., node classification, node clustering, metapath prediction, and visualization) under poisoning attack and evasion attack.
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 12218-12229
Where Am I From? Identifying Origin of LLM-generated Content
Liying Li, Yihan Bai, Minhao Cheng. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 12218-12229
Open AccessWhere Am I From? Identifying Origin of LLM-generated Content
Liying Li, Yihan Bai, Minhao Cheng. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 7518-7534
GuardEmb: Dynamic Watermark for Safeguarding Large Language Model Embedding Service Against Model Stealing Attack
Liaoyaqi Wang, Minhao Cheng. Findings of the Association for Computational Linguistics: EMNLP 2024. 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 7518-7534
Open AccessGuardEmb: Dynamic Watermark for Safeguarding Large Language Model Embedding Service Against Model Stealing Attack
Liaoyaqi Wang, Minhao Cheng. Findings of the Association for Computational Linguistics: EMNLP 2024. 2024.
- Open Access
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 4743-4755
Revisiting Personalized Federated Learning: Robustness Against Backdoor Attacks
In this work, besides improving prediction accuracy, we study whether personalization could bring robustness benefits to backdoor attacks. We conduct the first study of backdoor attacks in the pFL framework, testing 4 widely used backdoor attacks against 6 pFL methods on benchmark datasets FEMNIST and CIFAR-10, a total of 600 experiments. The study shows that pFL methods with partial model-sharing can significantly boost robustness against backdoor attacks. In contrast, pFL methods with full model-sharing do not show robustness. To analyze the reasons for varying robustness performances, we provide comprehensive ablation studies on different pFL methods. Based on our findings, we further propose a lightweight defense method, Simple-Tuning, which empirically improves defense performance against backdoor attacks. We believe that our work could provide both guidance for pFL application in terms of its robustness and offer valuable insights to design more robust FL methods in the future. We open-source our code to establish the first benchmark for black-box backdoor attacks in pFL: https://github.com/alibaba/FederatedScope/tree/backdoor-bench.
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 4743-4755
Open AccessRevisiting Personalized Federated Learning: Robustness Against Backdoor Attacks
In this work, besides improving prediction accuracy, we study whether personalization could bring robustness benefits to backdoor attacks. We conduct the first study of backdoor attacks in the pFL framework, testing 4 widely used backdoor attacks against 6 pFL methods on benchmark datasets FEMNIST and CIFAR-10, a total of 600 experiments. The study shows that pFL methods with partial model-sharing can significantly boost robustness against backdoor attacks. In contrast, pFL methods with full model-sharing do not show robustness. To analyze the reasons for varying robustness performances, we provide comprehensive ablation studies on different pFL methods. Based on our findings, we further propose a lightweight defense method, Simple-Tuning, which empirically improves defense performance against backdoor attacks. We believe that our work could provide both guidance for pFL application in terms of its robustness and offer valuable insights to design more robust FL methods in the future. We open-source our code to establish the first benchmark for black-box backdoor attacks in pFL: https://github.com/alibaba/FederatedScope/tree/backdoor-bench.
- Open Access
Stochastic Zeroth-order Optimization via Variance Reduction method
Derivative-free optimization has become an important technique used in machine learning for optimizing black-box models. To conduct updates without explicitly computing gradient, most current approaches iteratively sample a random search direction from Gaussian distribution and compute the estimated gradient along that direction. However, due to the variance in the search direction, the convergence rates and query complexities of existing methods suffer from a factor of $d$, where $d$ is the problem dimension. In this paper, we introduce a novel Stochastic Zeroth-order method with Variance Reduction under Gaussian smoothing (SZVR-G) and establish the complexity for optimizing non-convex problems. With variance reduction on both sample space and search space, the complexity of our algorithm is sublinear to $d$ and is strictly better than current approaches, in both smooth and non-smooth cases. Moreover, we extend the proposed method to the mini-batch version. Our experimental results demonstrate the superior performance of the proposed method over existing derivative-free optimization techniques. Furthermore, we successfully apply our method to conduct a universal black-box attack to deep neural networks and present some interesting results.Open AccessStochastic Zeroth-order Optimization via Variance Reduction method
Derivative-free optimization has become an important technique used in machine learning for optimizing black-box models. To conduct updates without explicitly computing gradient, most current approaches iteratively sample a random search direction from Gaussian distribution and compute the estimated gradient along that direction. However, due to the variance in the search direction, the convergence rates and query complexities of existing methods suffer from a factor of $d$, where $d$ is the problem dimension. In this paper, we introduce a novel Stochastic Zeroth-order method with Variance Reduction under Gaussian smoothing (SZVR-G) and establish the complexity for optimizing non-convex problems. With variance reduction on both sample space and search space, the complexity of our algorithm is sublinear to $d$ and is strictly better than current approaches, in both smooth and non-smooth cases. Moreover, we extend the proposed method to the mini-batch version. Our experimental results demonstrate the superior performance of the proposed method over existing derivative-free optimization techniques. Furthermore, we successfully apply our method to conduct a universal black-box attack to deep neural networks and present some interesting results.