16 July 2021
Artificial Intelligence and Machine Learningonline course
This course is designed to introduce students to the basic concepts of machine learning (ML) and artificial intelligence (AI) in a hands-on manner. We have developed the course in a self-contained manner with only basic knowledge of calculus and linear algebra required. Prior knowledge of machine learning and artificial intelligence is not essential.
The course will begin with a quick introduction to Python and the theoretical foundations of basic concepts in machine learning and artificial intelligence. Students will start with a simple linear regression example where they will derive and implement the gradient descent for a curve fitting problem and try to understand the concepts of loss function, regularization techniques, and bias-variance trade-off. Students will then be introduced to stochastic gradients descent and will implement stochastic gradient descent for regression using TensorFlow and Pytorch.
Students will design simple neural networks for MNIST classification and implement the full forward and backward pass for the training of the neural network. Following which students will be introduced to Convolutional Neural Networks and will implement MNIST classification with CNNs. Students will understand how Pytorch and TensorFlow handles the forward and backward pass during training. In the final part of the course, large scale problems of semantic segmentation, edge detection and metric learning will be implemented on AWS/ Google cloud.
As exercises for the course, the students will try to solve small scale practical problems of machine learning and artificial intelligence from diverse domains.
Dr Naeemullah Khan, Research Fellow, Lady Margaret Hall,
Postdoctoral Research Scientist, Department of Engineering, University of Oxford
The course is open to undergraduate and post-graduate students who want to get an introduction to Artificial Intelligence and Machine Learning.
In order to benefit from the course, knowledge of basic calculus and linear algebra is required.
To fully participate in the course all students will need to have proficiency in English to the following standards:
English language requirements for non-native English speaking students: Overall TOEFL score of 85; or IELTS score of 6.5 (no less than 6.0 in each component); or CET-4 at 550 or CET-6 at 520
There are no other formal academic entry criteria but we expect students to have a high level of commitment to their study.
Benefits / Outcomes - By the end of the course, the students will:
o Understand the theory of machine learning and artificial intelligence
o Know about ML and AI tools used in practice
o Know how to implement basic algorithms of AI and ML and train small networks for practical problems
o Be able to identify and use relevant AI and ML tools in their research
o Know how to implement and deploy ML and AI algorithms on AWS/Google cloud.
Programme structure / Teaching methods
This is an intensive hands-on course consisting of 10 full days of teaching and practical sessions. Every day will start with a 4-hour teaching session, from 9am-12pm UK time, where the instructor will explain theory and go over the exercises for each day. Following the instruction the students will be expected to work independently on their daily exercises for approximately 2 hours each day. The instructor will be available online during this time to support students in their independent work and to provide advice where needed.
Certificate of attendance
GBP 750: Early bird fee (ends 4 June 2021)
GBP 900: Standard fee (after 4 June 2021)