Strides in machine learning are some of the most cutting-edge technological advancements today. Steeped in the fields of neuroscience and artificial intelligence studies, it benefits from a continually growing body of knowledge of how the human brain receives and processes stimuli.
When applied to the juncture of technical innovation and art-making, machine learning explores the possibilities and limits of what a computer can do when it follows unusual objectives. Deep Dreaming is a term that refers to a strange process of object generation (fueled by computer vision) using a deep network, a technique from a Google researcher that has seen increasing attention since 2015 or so. An image is repeatedly filtered through a neural network to produce a completely new picture, which can be likened to a “hallucinatory” afterimage of the original.
Machine intelligence researcher and artist Parag Mital has worked on a plethora of interdisciplinary research projects, such as building computational models on audiovisual perception and augmented reality hallucinations. Mital demonstrates the Deep Dream technique in the video below.
For an informative, self-paced learning path in deep learning concepts and techniques, check out the 3-course Program Creative Applications of Deep Learning With TensorFlow. Join the first course for free below, or check out the first session here.
Creative Applications of Deep Learning With TensorFlow
Kadenze Academy