Deep learning uses big neural networks with several levels of processing units, Benefiting from advances in computing power and enhanced training techniques to learn elaborate designs in large amounts of data. Common applications contain image and speech recognition. For each doable motion, it may possibly estimate the "expected utility": the ut