We develop novel computational methods with the use ofmachine learning. Our approach improves the assignmentof different conformational states and enables effectiveintegration of diverse experimental data, while allowing usto make optimal use of both experimental insights andknowledge-based inferences