Qihang Chen
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
Qihang Chen has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
Bingrui Wu
NHC Key Laboratory of Glycoconjuates Research Department of Biochemistry and Molecular Biology School of Basic Medical Sciences Fudan University Shanghai 200032 P. R. China
1 shared publication- Chao Zhang
Beihang University
1 shared publication - Chaoyun Zhang
Microsoft, Beijing, China
1 shared publication Danfang Shi
NHC Key Laboratory of Glycoconjuates Research Department of Biochemistry and Molecular Biology School of Basic Medical Sciences Fudan University Shanghai 200032 P. R. China
1 shared publication- Dongmei Zhang
Microsoft Research Asia, Beijing, China
1 shared publication - Hao Fan
Microsoft, Suzhou, China
1 shared publication
Latest Publications
- Open Access
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 5116-5125
Root Cause Analysis for Microservice Systems via Hierarchical Reinforcement Learning from Human Feedback
In microservice systems, the identification of root causes of anomalies is imperative for service reliability and business impact. This process is typically divided into two phases: (i)constructing a service dependency graph that outlines the sequence and structure of system components that are invoked, and (ii) localizing the root cause components using the graph, traces, logs, and Key Performance Indicators (KPIs) such as latency. However, both phases are not straightforward due to the highly dynamic and complex nature of the system, particularly in large-scale commercial architectures like Microsoft Exchange. In this paper, we propose a new framework that employs Hierarchical Reinforcement Learning from Human Feedback (HRLHF) to address these challenges. Our framework leverages the static topology of the microservice system and efficiently employs the feedback of engineers to reduce uncertainty in the discovery of the service dependency graph. The framework utilizes reinforcement learning to reduce the number of queries required from O(N2) to O(1), enabling the construction of the dependency graph with high accuracy and minimal human effort. Additionally, we extend the discovered dependency graphs to window causal graphs that capture the characteristics of time series over a specified time period, resulting in improved root cause analysis accuracy and robustness. Evaluations on both real datasets from Microsoft Exchange and synthetic datasets with injected anomalies demonstrate superior performance on various metrics compared to state-of-the-art methods. It is worth mentioning that, our framework has been integrated as a crucial component in Microsoft M365 Exchange service.
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 5116-5125
Open AccessRoot Cause Analysis for Microservice Systems via Hierarchical Reinforcement Learning from Human Feedback
In microservice systems, the identification of root causes of anomalies is imperative for service reliability and business impact. This process is typically divided into two phases: (i)constructing a service dependency graph that outlines the sequence and structure of system components that are invoked, and (ii) localizing the root cause components using the graph, traces, logs, and Key Performance Indicators (KPIs) such as latency. However, both phases are not straightforward due to the highly dynamic and complex nature of the system, particularly in large-scale commercial architectures like Microsoft Exchange. In this paper, we propose a new framework that employs Hierarchical Reinforcement Learning from Human Feedback (HRLHF) to address these challenges. Our framework leverages the static topology of the microservice system and efficiently employs the feedback of engineers to reduce uncertainty in the discovery of the service dependency graph. The framework utilizes reinforcement learning to reduce the number of queries required from O(N2) to O(1), enabling the construction of the dependency graph with high accuracy and minimal human effort. Additionally, we extend the discovered dependency graphs to window causal graphs that capture the characteristics of time series over a specified time period, resulting in improved root cause analysis accuracy and robustness. Evaluations on both real datasets from Microsoft Exchange and synthetic datasets with injected anomalies demonstrate superior performance on various metrics compared to state-of-the-art methods. It is worth mentioning that, our framework has been integrated as a crucial component in Microsoft M365 Exchange service.
- Open Access
- 7 July 2022
- Published by Wiley in Advanced Science
- Vol. 9 (26) , e2202216
The Interaction between DNMT1 and High‐Mannose CD133 Maintains the Slow‐Cycling State and Tumorigenic Potential of Glioma Stem Cell
The quiescent/slow‐cycling state preserves the self‐renewal capacity of cancer stem cells (CSCs) and leads to the therapy resistance of CSCs. The mechanisms maintaining CSCs quiescence remain largely unknown. Here, it is demonstrated that lower expression of MAN1A1 in glioma stem cell (GSC) resulted in the formation of high‐mannose type N‐glycan on CD133. Furthermore, the high‐mannose type N‐glycan of CD133 is necessary for its interaction with DNMT1. Activation of p21 and p27 by the CD133–DNMT1 interaction maintains the slow‐cycling state of GSC, and promotes chemotherapy resistance and tumorigenesis of GSCs. Elimination of the CD133–DNMT1 interaction by a cell‐penetrating peptide or MAN1A1 overexpression inhibits the tumorigenesis of GSCs and increases the sensitivity of GSCs to temozolomide. Analysis of glioma samples reveals that the levels of high‐mannose type N‐glycan are correlated with glioma recurrence. Collectively, the high mannose CD133–DNMT1 interaction maintains the slow‐cycling state and tumorigenic potential of GSC, providing a potential strategy to eliminate quiescent GSCs.
- 7 July 2022
- Published by Wiley in Advanced Science
- Vol. 9 (26) , e2202216
Open AccessThe Interaction between DNMT1 and High‐Mannose CD133 Maintains the Slow‐Cycling State and Tumorigenic Potential of Glioma Stem Cell
The quiescent/slow‐cycling state preserves the self‐renewal capacity of cancer stem cells (CSCs) and leads to the therapy resistance of CSCs. The mechanisms maintaining CSCs quiescence remain largely unknown. Here, it is demonstrated that lower expression of MAN1A1 in glioma stem cell (GSC) resulted in the formation of high‐mannose type N‐glycan on CD133. Furthermore, the high‐mannose type N‐glycan of CD133 is necessary for its interaction with DNMT1. Activation of p21 and p27 by the CD133–DNMT1 interaction maintains the slow‐cycling state of GSC, and promotes chemotherapy resistance and tumorigenesis of GSCs. Elimination of the CD133–DNMT1 interaction by a cell‐penetrating peptide or MAN1A1 overexpression inhibits the tumorigenesis of GSCs and increases the sensitivity of GSCs to temozolomide. Analysis of glioma samples reveals that the levels of high‐mannose type N‐glycan are correlated with glioma recurrence. Collectively, the high mannose CD133–DNMT1 interaction maintains the slow‐cycling state and tumorigenic potential of GSC, providing a potential strategy to eliminate quiescent GSCs.
- 1 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Reliability Life Evaluation Model of Multifield Coupling Rotating Lip Seal based on Surface Simulation
Rotating lip seal as a common mechanical dynamic seal, mainly composed of a metal frame, a rubber sealing lip and a garter spring. Considering the complex microstructure of sealing lip and changing working conditions, this paper establishes a reliability life assessment model with multi-field coupling based on the simulation of rough surface. In the surface simulation of sealing lip, compared with the traditional Gaussian filter random surface generation method, fractal theory is applied to fit the complex surface more objectively and accurately. Multi-field coupling is also considered comprehensively, including fluid mechanics, elastic deformation and thermal. Through the cycle iteration of model parameters, performance degradation parameters of the lip seal are obtained to evaluate the reliability life, including friction torque, pumping rate and contact temperature. Finally, the parameters are measured by reliability test of rotary lip seal to verily the accuracy of the model.
- 1 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Reliability Life Evaluation Model of Multifield Coupling Rotating Lip Seal based on Surface Simulation
Rotating lip seal as a common mechanical dynamic seal, mainly composed of a metal frame, a rubber sealing lip and a garter spring. Considering the complex microstructure of sealing lip and changing working conditions, this paper establishes a reliability life assessment model with multi-field coupling based on the simulation of rough surface. In the surface simulation of sealing lip, compared with the traditional Gaussian filter random surface generation method, fractal theory is applied to fit the complex surface more objectively and accurately. Multi-field coupling is also considered comprehensively, including fluid mechanics, elastic deformation and thermal. Through the cycle iteration of model parameters, performance degradation parameters of the lip seal are obtained to evaluate the reliability life, including friction torque, pumping rate and contact temperature. Finally, the parameters are measured by reliability test of rotary lip seal to verily the accuracy of the model.