Summary:

Abstract: CHIPS was a 5 kton WC detector, constructed in 2019 to demonstrate a number of ideas associated with bringing the cost of neutrino detectors to a level where arrays of such detectors could be envisioned in order to potentially achieve greater precision on the neutrino oscillation parameters for a much lower cost. We will outline the cost-saving concepts and describe the construction process. Modern Machine Learning algorithms (specifically convolutional neural networks) were introduced to characterise raw neutrino events recorded within CHIPS. The performance gains from this new methodology were significant enough to alleviate the comparatively low density of instrumentation on the inner surface of the WC detector. We will detail how these algorithms were implemented and the specific challenges of applying them to WC detectors.