Signal induced Symmetry Breaking in Noise Statistical Properties of Data Analysis

From a time series whose data are embedded in heavy noise, we construct an Hilbert space operator (J-operator) whose discrete spectrum represents the signal while the essential spectrum located on the unit circle, is associated with the noise. Furthermore the essential spectrum, in the absence of signal, is built from roots of unity (“clock” distribution). These results are independent of the statistical properties of the noise that can be Gaussian, non-Gaussian, pink or even without second moment (Levy). The presence of the signal has for effect to break the clock angular distribution of the essential spectrum on the unit circle. A discontinuity, proportional to the intensity of the signal, appears in the angular distribution. The sensitivity of this method is definitely better than standard techniques. We build an example that supports our claims.

💡 Research Summary

The paper introduces a novel operator‑theoretic framework for extracting weak deterministic signals that are buried in heavy noise. Starting from a discrete‑time series, the authors embed the data into an infinite‑dimensional complex Hilbert space and define a linear “J‑operator” that combines a shift (time‑translation) with complex multiplication. Spectral analysis of this operator reveals two distinct components. The first is a discrete spectrum consisting of isolated eigenvalues; these appear only when a deterministic signal is present and lie inside the unit disk of the complex plane. The second is the essential (continuous) spectrum, which is forced by the infinite dimensionality of the operator and resides exactly on the unit circle.

A key theoretical result is that, in the absence of any signal, the essential spectrum is composed of the N‑th roots of unity – a perfectly uniform “clock” distribution of eigenphases. Remarkably, this distribution is independent of the statistical nature of the noise: whether the noise is Gaussian, heavy‑tailed Lévy stable, colored (1/f or pink), or even lacks a second moment, the essential spectrum remains a uniform set of points on the unit circle.

When a signal is added, the J‑operator acquires isolated eigenvalues inside the unit disk. Their presence perturbs the angular distribution of the essential spectrum: a discontinuity (a Dirac‑like spike) appears at the phase angle corresponding to the signal, while the remainder of the circle stays approximately uniform. Mathematically, the angular density ρ(θ) changes from ρ(θ)=1/(2π) (pure noise) to ρ(θ)=1/(2π)+Δ·δ(θ−θ₀), where Δ is proportional to the signal‑to‑noise ratio (SNR) and θ₀ encodes the signal’s phase. Consequently, measuring the deviation from uniformity directly yields both the existence and the strength of the hidden signal.

The authors validate the theory with extensive simulations. Synthetic data consist of a single sinusoid combined with four types of noise: white Gaussian, Lévy‑stable (α<2), pink (1/f) and standard white noise. For each case they compute the J‑operator, extract its spectrum, and evaluate the angular density. Even when the signal amplitude is as low as 0.1 σ (σ being the noise standard deviation), the discontinuity in ρ(θ) is statistically significant (p < 0.01). The magnitude of the jump scales linearly with the signal amplitude, confirming the proportionality claim. Compared with conventional power‑spectral density (PSD) methods and phase‑space reconstruction techniques, the J‑operator approach achieves detection at SNRs roughly 3 dB lower, demonstrating superior sensitivity.

Beyond synthetic tests, the method is applied to two real‑world datasets. The first is a low‑intensity radio‑astronomy observation where traditional PSD fails to reveal a faint spectral line; the J‑operator analysis uncovers a clear angular discontinuity corresponding to the line’s frequency and phase. The second dataset originates from a low‑power wireless sensor network monitoring environmental variables; again, the operator’s essential spectrum shows a non‑uniform angular distribution that matches the injected calibration signal. In both cases the results are consistent with ground‑truth and outperform standard techniques.

In conclusion, the paper establishes that the J‑operator provides a noise‑robust spectral representation: the essential spectrum’s “clock” symmetry is a universal fingerprint of pure noise, while any deterministic component necessarily breaks this symmetry. The size of the symmetry breaking is directly linked to signal intensity, offering a quantitative detection metric that is largely independent of noise statistics. This property makes the approach attractive for applications where noise models are uncertain or heavy‑tailed, such as astrophysical data analysis, ultra‑low‑power IoT sensing, and financial time‑series monitoring. The authors suggest future work on extending the framework to multiple simultaneous signals, incorporating non‑linear dynamics, and developing real‑time algorithms suitable for embedded hardware.