Improvement of machine learning-based vertex reconstruction for large liquid scintillator detectors with multiple types of PMTs

The precise vertex reconstruction for large liquid scintillator detectors is essential. A novel machine learning-based method was successfully developed to reconstruct an event vertex in JUNO. In this study, the performance of machine learning-based vertex reconstruction was further improved by optimizing the input images of neural networks. By separating the information of different types of PMTs and adding the information of the second hit of PMTs, the vertex resolution was improved by approximately 9.4 % at 1 MeV and 9.8 % at 11 MeV.