Since 2017, MIT has been organizing an online introductory course to Deep Learning (entirely in English) which, in addition to being aimed at students of the American institution, allows the registration of any user interested in the subject.
Now, finally, the website of this year's edition of the so-called 'MIT Introduction to Deep Learning' is available (although previous editions are also accessible online).
Since this past weekend, the slide and video of the first conference (given last Friday) have been available online ... even without prior registration, just by accessing the official website.
There, we can see how the course calendar foresees the availability of a new conference each week, until the twelfth and final that will take place on April 23. To this are added three 'software labs' whose source codes are also already available.
Course syllabus and requirements
The organizers note that the level of the course assumes prior knowledge in the field of calculus and linear algebra; in terms of programming, they state that "experience with Python is helpful, but not necessary."
The course agenda addresses aspects of deep learning such as computer vision or natural language processing, through reinforcement learning, generative models or the uses of deep learning in the field of biosanitary sciences.
"Students will gain a basic understanding of deep learning algorithms and gain practical experience in building neural networks using TensorFlow."
"The course concludes with a project proposal contest with input from staff and the sponsor panel."
Here's the first video lesson of the course:
Image | IBM Research