Causal Discovery in Structural VAR Models Under Equal Noise Variance

Causal discovery from multivariate time series is challenging when causal effects may occur both across time and within the same sampling interval. This issue is especially important in applications such as neuroscience, where the sampling rate may be coarse relative to the underlying dynamics and contemporaneous effects need not form an acyclic graph. We study causal discovery in linear Gaussian structural VAR models under an equal noise variance assumption, meaning that the structural noise terms have a common variance. Unlike the DAG-based cross-sectional equal noise variance setting, the time-series setting considered here does not generally yield point identification of a unique causal graph. Instead, multiple structural VAR parameterizations can induce the same stationary observed process law. We introduce a notion of observational equivalence tailored to this setting and show that the corresponding equivalence class is characterized by orthogonal transformations of the structural equations together with a global positive scale. This characterization leads to an equivalence-aware model discrepancy, the observational alignment discrepancy, which compares structural models modulo transformations that preserve the observed law. Building on this theory, we propose ENVAR, a sparsity-based procedure that searches over the induced observational equivalence class for a sparse normalized structural representative. We evaluate the proposed methodology on synthetic structural VAR data and on an fMRI dataset.

Machine Learning Based Framework for Estimation of Data Center Power Using Acoustic Side Channel

Data centers are high power consumers and the energy consumption of data centers keeps on rising in spite of all the efforts for increasing the energy efficiency. The need for energy-awareness in data centers makes the use of power modeling and estimation to be still a big challenge due to huge amount of uncertainty in this area. In this paper, a machine learning based method is proposed to approximately estimate the amount of power consumption by using acoustic side channel caused by fan in the fan-based cooling system in the server room. For doing so, frequency components of the acoustic signal, recorded by a microphone in the server room, is extracted, pre-processed, and fed to a Multi-Layer Neural-Network as an estimator. The proposed method performed well to estimate the power consumption, having more than 85 percent accuracy.