BPMN modelling of the fine-tuning process

The current BPMN diagram, depicting the fine-tuning process of a machine learning model. Fine-tuning diagram does contain security requirements for brevity due to repetition of applicable security requirements.

There are 3 main BPMN swim lanes, depicting:

1. "Machine learning training system" (top lane) – the system asset standing for the collection of tools, services and processes responsible for the training of a machine learning model.
2. "ML system input/API" (middle lane) – the system asset, representing an interface, which receives user input, transfers it for preprocessing and for inference to the model.
3. "Enabling software components" (bottom lane) - represents the corresponding system asset and the processes, which are undertaken with the asset. The components may include software frameworks, libraries, and runtime environments, utilized for model building, training and development.

The diagram represents the possible variation of fine-tuning, which takes preprocessed input query and utilizes it to update, re-train the model. The process starts with the submission of the input data through the input system. The input system exchanges the input data with the enabling software components for the input data preprocessing and model refinement. The enabling software components pass the preprocessed input data to the machine learning training system. The training system re-trains, tunes the machine learning model and returns it to the enabling components for deployment.