PhotonIDs: ML-Powered Photon Identification System for Dark Count Elimination

TL;DR

PhotonIDs addresses dark counts in SNSPD-based quantum networks by delivering an end-to-end system that processes full SNSPD waveforms in real time. It combines FPGA-based acquisition with a hybrid machine learning model that uses KDE-anchored pseudo-positions, CNN regression, and monotone calibration to distinguish photons from dark counts with high accuracy. Experimental results from a 20-km quantum link and an erbium-based emitter show substantial SNR improvements, including about $G\approx 31.2\times$ stream-level gain (≈$+14.9$ dB) and up to $98\%$ classification accuracy, illustrating practical impact for scalable, noise-resilient quantum communication. Overall, PhotonIDs provides a scalable pathway toward dark-count-free quantum networking at distance and across diverse photon sources.

Abstract

Reliable single photon detection is the foundation for practical quantum communication and networking. However, today's superconducting nanowire single photon detector(SNSPD) inherently fails to distinguish between genuine photon events and dark counts, leading to degraded fidelity in long-distance quantum communication. In this work, we introduce PhotonIDs, a machine learning-powered photon identification system that is the first end-to-end solution for real-time discrimination between photons and dark count based on full SNSPD readout signal waveform analysis. PhotonIDs ~demonstrates: 1) an FPGA-based high-speed data acquisition platform that selectively captures the full waveform of signal only while filtering out the background data in real time; 2) an efficient signal preprocessing pipeline, and a novel pseudo-position metric that is derived from the physical temporal-spatial features of each detected event; 3) a hybrid machine learning model with near 98% accuracy achieved on photon/dark count classification. Additionally, proposed PhotonIDs ~ is evaluated on the dark count elimination performance with two real-world case studies: (1) 20 km quantum link, and (2) Erbium ion-based photon emission system. Our result demonstrates that PhotonIDs ~could improve more than 31.2 times of signal-noise-ratio~(SNR) on dark count elimination. PhotonIDs ~ marks a step forward in noise-resilient quantum communication infrastructure.

Figures (17)

• Figure 1: An overview of the proposed hybrid machine learning powered photon identification system on the classification of photon and dark count. The system includes SNSPD setup, FPGA data acquisition platform, and the proposed hybrid ML model.
• Figure 2: Operational mechanism of SNSPD and readout signal with three phases of variation.
• Figure 3: Today's single photon detection system. The working principle is binary in nature, utilizing a predefined threshold as a comparator to count the number of detected events regardless of whether they are real photons or dark counts.
• Figure 4: Hardware platform of PhotonIDs based on the ZCU111 RFSoC board. The ARM CPU runs a Python-based PYNQ environment to configure the FPGA, which controls the ADCs and processes signals in real time.
• Figure 5: Event driven acquisition in PhotonIDs. The system transitions between IDLE, ARMED, TRIGGER, and INHIBITION states to capture valid waveforms only while suppressing continuous background data highlighted as red square .