Paper title: Guo, Zhen, Guofei Jiang, Haifeng Chen, and Kenji Yoshihira. "Tracking probabilistic correlation of monitoring data for fault detection in complex systems." In Dependable Systems and Networks, 2006. DSN 2006. International Conference on, pp. 259-268. IEEE, 2006. 1. Scope and relevance: Relevance to the REALM project (1=borderline 5=spot on): 3-4 Contribution to the community (# citations, work extended by others?): 61 What ML methods are used? Probabilistic modelling (GMM) and parameter estimation via EM What is the problem class / use case? (anomaly detection, regression, classification etc): Anomaly detection/fault detection/fault localization What is the application domain? Internet What type of data is studied? (time-series, steady-state, static/dynamic): Time-series, correlation between user requests and flow intensity of monitoring data. 2. Quality and scientific soundness: Clarity of the presentation. structure, is the problem well defined? The problem is to probabilistically correlate flow-intensities at multiple points in the network in order to detect and identify faulty states in the service/network. Are the methods well described/referenced? The authors provide a structured overview of how to exploit the severity in observed anomalies relative to a learned baseline model for the purpose of anomaly detection. Are the experiments repeatable/extendable? The estimated models have been tested against empirical models, but needs further evaluation. Some examples of anomaly detection are provided but needs further evaluation as well. Are data sets publicly available? The experiments are based on synthetically generated data using a client emulation approach in Java. Are alternative methods evaluated? This is only preliminary results that will be/has been extended in other papers.