Zizhao Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zizhao Zhang has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Yujie Xiao
University of California, Berkeley, USA
3 shared publications - Chi Zhang
Northeastern University, USA
2 shared publications - Haotian Zhu
New York University, USA
2 shared publications - Jinming Li
Georgia Institute of Technology, USA
2 shared publications - Sijia Li
University of Michigan, USA
2 shared publications - Ao Zhu
University of Pennsylvania, USA
1 shared publication
Latest Publications
- Open Access
- 27 February 2026
- Published by MDPI AG in Preprints.org
Deep Learning Approach for Protocol Anomaly Detection Using Status Code Sequences
This paper addresses the limitations of traditional protocol anomaly detection methods in handling dynamic state changes and unstructured behaviors. A deep protocol anomaly detection algorithm based on status code sequence modeling is proposed. The method uses the status codes returned during protocol communication as the core input. A state embedding layer is employed to transform discrete status codes into continuous vector representations. A gated recurrent unit (GRU) is then used to capture temporal dependencies and behavior patterns within the status sequence. Based on this structure, the model integrates sequence reconstruction and contrastive learning mechanisms. Reconstruction error is used to characterize the normal distribution of sequences. Contrastive loss is introduced to enhance the model's ability to distinguish abnormal states. The paper also conducts a series of sensitivity experiments on key hyperparameters, including the number of hidden units, activation functions, and the temperature coefficient. These experiments verify the influence of each module on the overall performance of the model. The proposed method achieves superior performance across multiple evaluation metrics. It effectively identifies potential abnormal behaviors at the protocol layer. The method also demonstrates good generalizability and strong detection capability.
- 27 February 2026
- Published by MDPI AG in Preprints.org
Open AccessDeep Learning Approach for Protocol Anomaly Detection Using Status Code Sequences
This paper addresses the limitations of traditional protocol anomaly detection methods in handling dynamic state changes and unstructured behaviors. A deep protocol anomaly detection algorithm based on status code sequence modeling is proposed. The method uses the status codes returned during protocol communication as the core input. A state embedding layer is employed to transform discrete status codes into continuous vector representations. A gated recurrent unit (GRU) is then used to capture temporal dependencies and behavior patterns within the status sequence. Based on this structure, the model integrates sequence reconstruction and contrastive learning mechanisms. Reconstruction error is used to characterize the normal distribution of sequences. Contrastive loss is introduced to enhance the model's ability to distinguish abnormal states. The paper also conducts a series of sensitivity experiments on key hyperparameters, including the number of hidden units, activation functions, and the temperature coefficient. These experiments verify the influence of each module on the overall performance of the model. The proposed method achieves superior performance across multiple evaluation metrics. It effectively identifies potential abnormal behaviors at the protocol layer. The method also demonstrates good generalizability and strong detection capability.
- Open Access
- 29 January 2026
- Published by MDPI AG in Preprints.org
Deep Learning-Based Uncertainty-Driven Robust Time Series Forecasting for Backend Service Metrics
This study addresses the challenge of forecasting backend service metrics in highly dynamic and disturbed environments and proposes a robust time series prediction framework based on uncertainty estimation. The method first applies a multi-scale encoding structure to decompose and represent service metric sequences, allowing the model to capture mixed characteristics of trends, periodic patterns, and noise disturbances. It then constructs a cross-sequence dependency modeling module to describe structural coupling among multiple metrics so that the model can identify implicit interaction patterns across different service nodes in backend systems. On this basis, the framework outputs both the mean and variance of future metrics through distributed prediction heads, enabling joint characterization of predictions and their associated risk ranges. A noise response mechanism is incorporated to enhance model stability under anomalous disturbances. To improve the controllability of the overall prediction process, a risk adjustment module is designed to transform uncertainty signals into actionable adjustment factors, allowing the model to maintain consistent outputs in high volatility environments. The experiments evaluate hyperparameter sensitivity, environmental sensitivity, and data sensitivity. The results show that the proposed framework maintains high stability and robustness under multiple source disturbances and effectively suppresses error propagation caused by anomalous samples, structural changes, and system scaling. By jointly modeling point predictions and risk estimation, the study verifies the effectiveness of an uncertainty-driven robust forecasting framework for backend metric management and system reliability, providing a scalable technical path for data-driven analysis in complex systems.
- 29 January 2026
- Published by MDPI AG in Preprints.org
Open AccessDeep Learning-Based Uncertainty-Driven Robust Time Series Forecasting for Backend Service Metrics
This study addresses the challenge of forecasting backend service metrics in highly dynamic and disturbed environments and proposes a robust time series prediction framework based on uncertainty estimation. The method first applies a multi-scale encoding structure to decompose and represent service metric sequences, allowing the model to capture mixed characteristics of trends, periodic patterns, and noise disturbances. It then constructs a cross-sequence dependency modeling module to describe structural coupling among multiple metrics so that the model can identify implicit interaction patterns across different service nodes in backend systems. On this basis, the framework outputs both the mean and variance of future metrics through distributed prediction heads, enabling joint characterization of predictions and their associated risk ranges. A noise response mechanism is incorporated to enhance model stability under anomalous disturbances. To improve the controllability of the overall prediction process, a risk adjustment module is designed to transform uncertainty signals into actionable adjustment factors, allowing the model to maintain consistent outputs in high volatility environments. The experiments evaluate hyperparameter sensitivity, environmental sensitivity, and data sensitivity. The results show that the proposed framework maintains high stability and robustness under multiple source disturbances and effectively suppresses error propagation caused by anomalous samples, structural changes, and system scaling. By jointly modeling point predictions and risk estimation, the study verifies the effectiveness of an uncertainty-driven robust forecasting framework for backend metric management and system reliability, providing a scalable technical path for data-driven analysis in complex systems.
- Open Access
- 31 December 2025
- Published by MDPI AG in Preprints.org
Memory-Driven Agent Planning for Long-Horizon Tasks via Hierarchical Encoding and Dynamic Retrieval
This study addresses the challenge of long-term dependency modeling in agent behavior planning for long-horizon tasks and proposes a memory-driven agent planning framework. The method introduces hierarchical memory encoding and dynamic memory retrieval structures, enabling the agent to selectively retain and effectively utilize historical information across multiple time scales, thereby maintaining policy stability and goal consistency in complex dynamic environments. The core idea is to construct an interaction mechanism between short-term and long-term memory, where attention-guided retrieval integrates historical experience with current perception to support continuous planning and decision optimization in long-term tasks. The proposed framework consists of four key modules: perception input, memory encoding, state updating, and behavior generation, forming an end-to-end task-driven learning process. Experimental evaluations based on success rate, average planning steps, memory consistency score, and policy stability demonstrate that the proposed algorithm achieves superior performance in long-term task scenarios, effectively reducing planning redundancy and improving strategy coherence and task efficiency. The results confirm that the memory-driven mechanism provides a novel theoretical foundation and algorithmic framework for developing long-term task agents, establishing a solid basis for adaptive decision-making and continuous planning in complex environments.
- 31 December 2025
- Published by MDPI AG in Preprints.org
Open AccessMemory-Driven Agent Planning for Long-Horizon Tasks via Hierarchical Encoding and Dynamic Retrieval
This study addresses the challenge of long-term dependency modeling in agent behavior planning for long-horizon tasks and proposes a memory-driven agent planning framework. The method introduces hierarchical memory encoding and dynamic memory retrieval structures, enabling the agent to selectively retain and effectively utilize historical information across multiple time scales, thereby maintaining policy stability and goal consistency in complex dynamic environments. The core idea is to construct an interaction mechanism between short-term and long-term memory, where attention-guided retrieval integrates historical experience with current perception to support continuous planning and decision optimization in long-term tasks. The proposed framework consists of four key modules: perception input, memory encoding, state updating, and behavior generation, forming an end-to-end task-driven learning process. Experimental evaluations based on success rate, average planning steps, memory consistency score, and policy stability demonstrate that the proposed algorithm achieves superior performance in long-term task scenarios, effectively reducing planning redundancy and improving strategy coherence and task efficiency. The results confirm that the memory-driven mechanism provides a novel theoretical foundation and algorithmic framework for developing long-term task agents, establishing a solid basis for adaptive decision-making and continuous planning in complex environments.
- Open Access
- 30 December 2025
- Published by MDPI AG in Preprints.org
Unsupervised Anomaly Detection in Cloud-Native Microservices via Cross-Service Temporal Contrastive Learning
We address a key challenge in cloud native microservice anomaly detection. Existing methods struggle to characterize cross-service dependency propagation and complex temporal dynamics. We propose a temporal representation framework based on cross-service contrastive learning. The method first constructs multi-scale views from service call relations and historical runtime data. It embeds system-level collaborative behaviors into temporal window representations. This yields structure-enhanced input features. It then applies a dual view contrastive encoding strategy with weak augmentation across different time windows. This strategy pulls normal collaborative patterns closer and separates abnormal deviation patterns in the feature space. It improves sensitivity and robustness. The framework further performs unsupervised anomaly scoring through feature aggregation and service-level prototype construction. This avoids extra labeling costs and reduces dependence on manual thresholds. To evaluate the framework, we conduct systematic validation using metrics, logs, and distributed traces from a real open source microservice benchmark system. We also perform multi-dimensional hyperparameter sensitivity analyses across learning rate, batch size, time window length, and the temperature coefficient of the contrastive loss. These analyses reveal stability and adaptability under different dynamic conditions. In comparison with existing baseline models, the method achieves higher accuracy, stronger recall, and better overall performance. This confirms the advantage of cross-service contrastive strategies for capturing anomaly propagation patterns and structural characteristics. Overall, the proposed unified structure of temporal representations and contrastive learning mechanism enables effective unsupervised anomaly detection in collaborative microservice environments. The framework shows strong discriminative capacity and robustness under complex system operations. It provides a reproducible technical solution for improving intelligent operations in cloud native systems.
- 30 December 2025
- Published by MDPI AG in Preprints.org
Open AccessUnsupervised Anomaly Detection in Cloud-Native Microservices via Cross-Service Temporal Contrastive Learning
We address a key challenge in cloud native microservice anomaly detection. Existing methods struggle to characterize cross-service dependency propagation and complex temporal dynamics. We propose a temporal representation framework based on cross-service contrastive learning. The method first constructs multi-scale views from service call relations and historical runtime data. It embeds system-level collaborative behaviors into temporal window representations. This yields structure-enhanced input features. It then applies a dual view contrastive encoding strategy with weak augmentation across different time windows. This strategy pulls normal collaborative patterns closer and separates abnormal deviation patterns in the feature space. It improves sensitivity and robustness. The framework further performs unsupervised anomaly scoring through feature aggregation and service-level prototype construction. This avoids extra labeling costs and reduces dependence on manual thresholds. To evaluate the framework, we conduct systematic validation using metrics, logs, and distributed traces from a real open source microservice benchmark system. We also perform multi-dimensional hyperparameter sensitivity analyses across learning rate, batch size, time window length, and the temperature coefficient of the contrastive loss. These analyses reveal stability and adaptability under different dynamic conditions. In comparison with existing baseline models, the method achieves higher accuracy, stronger recall, and better overall performance. This confirms the advantage of cross-service contrastive strategies for capturing anomaly propagation patterns and structural characteristics. Overall, the proposed unified structure of temporal representations and contrastive learning mechanism enables effective unsupervised anomaly detection in collaborative microservice environments. The framework shows strong discriminative capacity and robustness under complex system operations. It provides a reproducible technical solution for improving intelligent operations in cloud native systems.
- Open Access
- 29 December 2025
- Published by MDPI AG in Preprints.org
TrustOrch: A Dynamic Trust-Aware Orchestration Framework for Adversarially Robust Multi-Agent Collaboration
Multi-agent systems (MAS) have emerged as a critical paradigm for distributed problem-solving in complex environments. However, their deployment in mission-critical applications faces significant challenges regarding trust, security, and adversarial robustness. This paper presents TrustOrch, a novel dynamic trust-aware orchestration framework designed to enhance the resilience of multi-agent collaboration against adversarial attacks. TrustOrch introduces five key innovations: (1) a dynamic trust assessment mechanism that evaluates agent reliability in real-time using multi-dimensional metrics, (2) an adversary-aware orchestration strategy combining reinforcement learning and game theory to detect and mitigate prompt injection attacks, (3) an adaptive collaboration topology that dynamically adjusts agent communication structures based on task complexity and trust levels, (4) explainable decision tracing for complete audit chains, and (5) a layered security architecture leverag- ing blockchain technology for decentralized trust verification. Our experimental evaluation demonstrates that TrustOrch re- duces collision rates by 62%, achieves 91.7% robustness under adversarial attacks, and reduces communication overhead by 39.8% compared to baseline approaches. The framework achieves robust performance under various adversarial scenarios while maintaining transparency and regulatory compliance, making it particularly suitable for deployment in high-risk domains such as finance, healthcare, and autonomous systems.
- 29 December 2025
- Published by MDPI AG in Preprints.org
Open AccessTrustOrch: A Dynamic Trust-Aware Orchestration Framework for Adversarially Robust Multi-Agent Collaboration
Multi-agent systems (MAS) have emerged as a critical paradigm for distributed problem-solving in complex environments. However, their deployment in mission-critical applications faces significant challenges regarding trust, security, and adversarial robustness. This paper presents TrustOrch, a novel dynamic trust-aware orchestration framework designed to enhance the resilience of multi-agent collaboration against adversarial attacks. TrustOrch introduces five key innovations: (1) a dynamic trust assessment mechanism that evaluates agent reliability in real-time using multi-dimensional metrics, (2) an adversary-aware orchestration strategy combining reinforcement learning and game theory to detect and mitigate prompt injection attacks, (3) an adaptive collaboration topology that dynamically adjusts agent communication structures based on task complexity and trust levels, (4) explainable decision tracing for complete audit chains, and (5) a layered security architecture leverag- ing blockchain technology for decentralized trust verification. Our experimental evaluation demonstrates that TrustOrch re- duces collision rates by 62%, achieves 91.7% robustness under adversarial attacks, and reduces communication overhead by 39.8% compared to baseline approaches. The framework achieves robust performance under various adversarial scenarios while maintaining transparency and regulatory compliance, making it particularly suitable for deployment in high-risk domains such as finance, healthcare, and autonomous systems.