Machine Learning Enhances Detection of Double-\Lambda Hypernuclear Events

A new method has been developed to detect double-\Lambda hypernuclear events in nuclear emulsions using machine learning techniques. The research, conducted by Yan He and 19 co-authors, utilizes an object detection model known as Mask R-CNN, which was trained on images generated through Monte Carlo simulations and enhanced using image processing and generative adversarial networks.

The model achieved a detection efficiency of 93.9% for (^{6}\Lambda\Lambda\text{He}) events and 81.5% for (^{5}\Lambda\Lambda\text{H}) events. Notably, it successfully identified the Nagara event, the only uniquely identified (^{6}\Lambda\Lambda\text{He}) event reported to date. The model also demonstrated effective segmentation of event topologies.

In a significant application of this method, researchers analyzed 0.2% of the entire emulsion data from the J-PARC E07 experiment. This analysis led to the detection of six new candidates for double-\Lambda hypernuclear events, suggesting that over 2000 such events may be present in the complete dataset. The method is particularly valuable as it reduces the time required for manual visual inspection of nuclear emulsion sheets by a factor of 500, thereby enhancing the efficiency of data analysis in high-energy physics experiments.

This research is detailed in the paper titled "A novel machine learning method to detect double-\Lambda hypernuclear events in nuclear emulsions" by Yan He et al., available on arXiv at arXiv:2409.01657. The findings could have significant implications for future studies in nuclear physics and the understanding of hypernuclear structures.