Zhen Han Si

Title: Zhen Han Si and Causality Detection in Telemetry Data.

Introduction

Zhen Han Si is an inventor associated with Milpitas, California, and Palo Alto Networks, Inc. His patent work addresses automated analysis of telemetry data to help identify potential causes of unusual events.

Latest Patents

The provided information lists one patent for Si: “Causality Detection for Outlier Events in Telemetry Metric Data.” The invention uses machine-learning ensembles combining an autoencoder and an attention mechanism to identify causal relationships between outlier events in telemetry metrics. It first detects outliers across telemetry data using unsupervised learning models. It then analyzes pairs of metrics, inferring a causal relationship when the ensemble produces a sufficiently high loss value for sequences from both metrics. Attention-mechanism values can also identify timestamps in one metric that are associated with an outlier event in another.

Career Highlights

Si’s listed patent focuses on applying machine learning to root-cause analysis in cloud telemetry. The approach is designed to help automate the investigation of anomalous events by connecting related metric sequences and highlighting relevant points in time. The supplied information associates his work with Palo Alto Networks, Inc.

Collaborations

The listed coworkers associated with Si include Claudionor Jose Nunes Coelho, Jr. and Viswesh Ananthakrishnan. Their names are provided as professional collaborators, though the available details do not specify their roles in the patent.

Conclusion

Zhen Han Si’s listed invention applies autoencoder and attention-based machine learning to detect causal links among outlier telemetry events, supporting more automated root-cause analysis.