Deep learning is widely used in many market segments ranging from mobile devices to supercomputers. Recently different SW packages as well as special HW accelerators were developed to support deep learning. The course will focus on algorithms, programming languages and new SW/HW interfaces that aim to allow execution of deep learning algorithms in a productive and efficient way.
In this course we will deep dive to the cutting-edge research in Deep Learning applications and optimizations in various fields.During the course we will cover the theory and practice of neural network models (including: convolution neural networks,recurrent neural networks, long-short-term-memory models, variation auto encoders , deep reinforcement models such as policy gradients and q-learning).In addition we will cover parallel hardware architectures for efficient deep learning computation such as GPUs and dedicated hardware platforms
Course grade will be based on HW assignments and a final course project with lots of programming. Submissions will be done in pairs.
Course grade will be based on HW assignments and a final course project with lots of programming. Submissions will be done in pairs.
In order to enroll the course please submit application form (the file is attached under syllabus tab) and your CV including grade sheet to Chaim Baskin
uchaimbaskinatcs.technion.ac.il
Application deadline is 20.01.2018
Credit points: 3.0
Course will be taught in English.