Quantum Gas Analysis and Inference (Q-GAIN): A Python Package for Machine Learning and Physically Informed Analysis Applications

Q-GAIN, or Quantum Gas Analysis and Inference, is an evolution of the Soliton Detector (SolDet) Python library. The SolDet library was originally developed to unify the common task of feature identification and tracking in systems studying ultra-cold atoms into one cohesive framework. Q-GAIN takes this further by generalizing this concept into a growing modular architecture that can support a range of machine learning (ML) and conventional analysis techniques not constrained to any particular field of study. The library is applicable to any data suitable for ML tasks and emphasizes a workflow of tagging features with ML tools and applying conventional analysis techniques to the filtered data. It is designed to be modular in nature, allowing for the addition and removal of tools to meet the user's needs, and currently includes classification, object detection, and physically informed analysis methods for feature detection in absorption images of Bose-Einstein condensates. We demonstrate this with various applications that highlight how Q-GAIN is used to help explore and analyze experimental data.