About the course. Advanced Deep Learning is a 5-month online course. Deep learning is the technology which underpins the revolution in artificial intelligence. The need for experts in this field has never been larger. However, deep learning is complex, and the sheer volume of innovations can make it hard to focus on the skills that matter most. This course breaks down advanced topics step by step, ensuring you stay ahead in one of the fastest-growing fields in technology.
Who should apply
Apply if you meet one of the following descriptions:
Experienced machine learning industry professionals who want to delve deeper into advanced machine learning.
Those preparing to apply for an MSc/PhD programme in machine learning or computer science and are looking to build a solid foundation in deep learning.
Technical researchers applying machine learning methods in their work but desiring to go one step ahead.
Prerequisites. This course has prerequisites. You are expected to have taken classes or have a good foundation in the following: Statistics, Linear algebra, Calculus, Python. You can learn everything else on the go.
What you will learn
By the end of this course, you will:
For each topic, we will be analyzing how the theory connects to practical implementations and simple intuitions. The full syllabus is available HERE.
Course structure
Delivery format. There is one live lecture, and seminar each week conducted online. During the lectures, we will cover theory and concepts. During the seminar, we will do hands-on implementations and coding. There will be optional supplementary readings each week.
Class times. Tuesday or Wednesday evenings at 5:30 PM GMT+2 (final schedule to be confirmed after The course is divided into two parts. All students will begin with Part I. Only those who successfully complete Part I can move onto Part II).
Platform. All sessions are conducted via Zoom with materials available uploaded to an online learning platform.
About the Instructor
Paulius Rauba
Paulius is a machine learning scientist with extensive experience across academia, industry, and international organizations. He is currently pursuing a Ph.D. in Machine Learning at the University of Cambridge while serving as a Lecturer at ISM University of Management and Economics.
With a strong background in deep learning, generative AI, statistical modeling, and autonomous machine learning systems, Paulius has been actively engaged in AI research and development for over six years. His expertise spans data science, management consulting, and corporate banking, and he has contributed as an Expert for the European Commission on AI implementation.
Paulius holds a BSc from ISM University of Management and Economics and an MSc from the University of Oxford, where he was an Oxford Shirley Scholar. Passionate about teaching, he excels at making complex ideas accessible and is committed to helping others advance their careers in AI. Proficient in multiple languages, including Lithuanian, English, and French, he also has strong technical skills in Python, R, PyTorch, Keras, PySpark, and SQL.
Contact information for questions:
Curriculum Overview
The course is divided into two parts. All students will begin with Part I. Only those who successfully complete Part I can move onto Part II.
Foundations of Advanced Deep Learning. We’ll kick off with foundations that underpin everything: probabilities, information theory, and multivariate gaussian distributions and their properties. We’ll then move onto core topics in neural networks: maximum likelihood, decision theory, and properties of deep networks. We will learn how to use these principles for designing novel deep learning architectures.
Frontiers of Advanced Deep Learning. Building on part I, we will focus on modern approaches. These include attention-based mechanisms and associated transformer architectures, sampling methods (MCMC, Langevin sampling), latent variable modeling methods, probabilistic modeling (ELBO and variational inference methods), normalizing flows, and diffusion models.
Find the full syllabus of Advanced Deep Learning Syllabus HERE.
Application Process
Application review. After your application is received, you will receive an e-mail about whether you are moving forward in the selection process. You will receive an e-mail regardless of the outcome.
Intro call. If you’re successful, you will be invited to have a 15-minute introductory call with the lecturer, Paulius Rauba. This call has two goals: (i) to ensure you have the prerequisites of the course and (ii) to answer any questions you might have. You will be notified whether you are accepted into the course in 1-2 days after the call.
Payment and details. If you’re successful in the second stage, you will receive a message with administrative information about the course together with payment details. When we receive your payment for the course—you are officially on board.
Course starts. The course will begin on April 8th (Tuesday) or April 9th (Wednesday) at 5.30pm GMT +2. The exact start date will be confirmed two weeks before the course starts. There will be a lecture and a seminar held on the same day each week.
FAQs
Will the course be online or in-person?
The course will be conducted fully online.
Are the payments for Part I and Part II made separately?
Yes. You first pay for Part I. When Part I is finished (after 10 weeks), you will have the option to continue to Part II. The payment can be made then.
Is there a refund policy?
Yes. You can ask for a refund within 5 days after the first class/seminar.
Are there discounts for students?
No. There is one price for everyone.
Will the classes be recorded?
Yes. The classes will be uploaded to the appropriate e-learning platform together with the material (class notes, jupyter notebooks, and weekly readings).
How much knowledge of calculus, linear algebra, and python is required?
To feel comfortable in the course, you should know standard multivariable calculus (integration, derivatives). You should not feel afraid when you see an integral. For linear algebra, you should know how to multiply matrices, what vectors are, as well as basic concepts such as determinants. For Python, you should be comfortable working with Python classes and have knowledge of pandas, numpy, scikit-learn. It’s fantastic if you have experience with PyTorch but it’s okay if you do not.
Are there project assignments?
There will be a couple of major assignments that are optional. These project assignments will be focused on developing deep learning methods to solve data-intensive problems.
Do I need to have access to servers/compute?
No. Just have your own computer to work with.
What is the duration of a class and seminar?
Each seminar and lecture is about 1.5 hours each. Therefore, there is 3 hours of teaching/work done together per week.
How much time do I need to commit weekly?
There will be 3 hours of live sessions. Any additional time preparing is purely your own. I expect some students to study for > 10 hours each week and some to only come to the lectures and seminars.