A brain-like processor for knowledge-based reasoning

In current technology, music players are strictly dependent on the user´s input to play, repeat or discard certain songs. These tasks are performed according to the user´s preferences and indications on the system. In this sense, they are passive and the user´s preferences are recorded by evaluating his/her clicks which may also be unintentional. This paper presents a music preference indicator using a brain-like microprocessor, allowing the music player to actually “think” and develop preferences. The indicator senses the interest level of a user when listening to certain music genres. In this indicator system, the Q sensor was intended to use to retrieve the data that represents the reaction from the user. The audio classification software is responsible for obtaining useful data from the recorded sound wave. Combination of these data in neural networks results in pattern recognition units (PRUs). These PRU´s are connected together on an FPGA to form the brain-like architecture. Overall, this system has smart features to detect the user´s music preference and suggest similar songs and genres without the need to tell the music player.