Abstract: A method for diagnostic and condition monitoring of a system includes receiving data from one or more sensors, the data associated with the system; generating an audio feature based on the data; inputting the audio feature into a neural network model; and receiving one or more attribute predictions and a state prediction from the neural network model. In some embodiments, the monitored system is a vehicle and the one or more sensors are vibroacoustic sensors.