Journeys is a predictive inference and comprehension platform, it aims to make the process of gaining inferences from unstructured multi-modal data (i.e. text, image, video) easier for the novice by training machine learning models on the fly and automatically recognizing the intent of the user query
The overview of the entire process is illustrated in the figure, it involves a series of steps starting with query parsing, intent recognition, query parameterization, gathering relevant data, training ML models (using AutoML) followed by evaluating and interpreting the models (using model agnostic interpretation algorithms)
An example of the system works is elucidated in the figure given below -
