Deep Learning Lab
The work in this lab is characterized by engineering and training neural networks for various application domains and sensors.
The field of Deep Learning represents a variety of methods and applications based on building machine learning models consisting of multiple levels of composition. Fueled by the latest advances in computational capabilities and availability of data, such methods demonstrate themselves to be very efficient in different kinds of tasks: computer vision, speech recognition, natural language processing, and many others. Whether it’s unlocking your smartphone with a face detection algorithm, or using a voice assistant to search for restaurants — deep learning application works for you behind the curtains. Deep Learning can already be considered one of the biggest advances in Artificial Intelligence. At AGT we use Deep Learning technologies throughout most of our machine learning pipeline: from CNNs in image recognition, RNNs in audio recognition, LSTMs in action and activity recognition from wearables, to GANs that provide us new ways of simulating data.
From our lab team member:
“I joined AGT as a Data Scientist after finishing my PhD. Deep Learning brought artificial intelligence to a whole new level. At AGT I have a playground full of the latest methods in this fast-changing research field that allows me to solve some of the most complex classification tasks. It is the mixture of a strong connection to the scientific community and the experience of applying the latest neural network architectures to exciting consumer problems that is inspiring.”