The thesis describes the development of a web-based system for managing and analyzing machine learning experiments. The system allows dividing experiments in one or more tasks with mutual dependencies, comparing and displaying the results. Unlike existing systems and services, the presented system emphasizes the centralized organization and management of experiments, which is separate from the execution of experiments, which is not the subject of this thesis. The core of the system is a REST interface, intended for both communication with the user interface and with the workers performing the experiments.
|
