These tools and metrics are designed to help AI actors develop and use trustworthy AI systems and applications that respect human rights and are fair, transparent, explainable, robust, secure and safe.
NPSS can be used to monitor and validate the consistency and reliability of AI-generated outputs (e.g., motion predictions) under various conditions. By quantifying how closely predictions match expected spectral characteristics, it helps detect performance degradation or failures, thereby supporting the robustness objective. However, its connection is indirect and limited to specific technical contexts.The Normalized Power Spectrum Similarity (NPSS) is a metric used to compare the similarity between two power spectra. It is calculated by first normalizing both spectra to have unit energy, and then computing the cross-spectral density between the two normalized spectra. The NPSS is the square of the magnitude of this cross-spectral density, which ranges from 0 (completely dissimilar) to 1 (identical).
In summary, the NPSS metric measures how similar two power spectra are, after normalizing their energy, and is calculated using the cross-spectral density between the two spectra.
About the metric
You can click on the links to see the associated metrics
Objective(s):
Purpose(s):
Lifecycle stage(s):
Target users:





























