Detecting entangled photons is critical for quantum technologies but challenging due to weak signals, noise, and rarity. This study proposes a fast, unsupervised anomaly detection algorithm using a multivariate Gaussian distribution to identify potential entangled photon events. Applied to data from spontaneous parametric down-conversion (SPDC) and silicon photomultiplier (SiMP) detectors, it successfully detected rare correlated events, offering a simple, efficient tool for quantum optics.

Entangled photon detection is essential for advancements in quantum communication, cryptography, and fundamental quantum mechanics experiments. This study introduces a novel application of unsupervised machine learning for identifying potential entangled photon events by analyzing voltage signals recorded from Silicon Multiplier Amplified Detectors (SiMPs). By framing photon detection as an anomaly detection problem, we employ the Isolation Forest (iForest) algorithm to isolate rare and distinctive signal patterns within large, noisy datasets without requiring labeled training data. This is the first application of iForest in the context of entangled photon detection. The method enables automated identification of anomalous events exhibiting time correlations across multiple measurement channels, offering a scalable and computationally efficient solution for real-time processing of experimental data in quantum optics.

This paper presents research on software-based tools for the semi-automated collection and classification of scientific papers focusing on Quantum Optics research. The tool integrates keyword-based search, Portable Document Format extraction, and parallel processing using “term frequency-inverse document frequency” and the all-MiniLM-L6-v2 model for semantic analysis. It generates numerical similarity estimates, enabling efficient navigation and prioritization of sources. The modular design allows flexible integration of similarity algorithms. Despite challenges with dynamic anti-scraping mechanisms, the tool demonstrates significant potential in streamlining literature reviews. Future improvements include advanced NLP techniques and addressing ethical considerations to enhance accuracy and compliance.