DS409
Deep Learning in Applications
Faculty Profiles
Anastasia Ianina
Research Scientist at Meta Reality Labs
Michael Diskin
Lead LLM Research Engineer at Wildberries
Course length
Duration
Total hours
Credits
Language
Course type
Fee for single course
Fee for degree students
Skills you’ll learn
Overview
State-of-the-art approaches in different domains of artificial intelligence are based on deep learning techniques (e.g. computer vision, natural language processing, reinforcement learning, etc.). Deep neural architectures show great potential and promise even better results, so now is definitely the time to explore this field.
In this course we will start from the basics and rapidly dive into the latest results in Deep Learning, focusing on the NLP and RL domains. This course focuses both on practical skills and theoretical background to provide the students with thorough theoretical knowledge and ability to work on their own in the Deep Learning area.
This course accompanies the Machine Learning course (Module 6).
Programming assignments will be implemented in Python 3. The PyTorch framework will be used for Deep Learning practice.
Learning highlights
- Learn to apply Deep Learning techniques in practice
- Get familiar with both fundamental and most recent approaches in Natural Language Processing (NPL) and Reinforcement Learning
- Learn about large language models (LLMs) and how to apply them to different products.
- Get ready to face real-world problems and to apply the Deep Learning techniques to them
- Gain essential experience with the PyTorch framework
Course outline
15 classes
Natural Language Processing intro
Main problems in NLP. Text classification and generation. Deep learning techniques in NLP. Regularisation in DL recap. CNN approach to context analysis. Word embeddings recap.
Neural Machine Translation
Machine translation and neural machine translation. Encoder-decoder architecture, sequential modelling.
Attention in Encoder-Decoder Architecture
Encoder-decoder architecture bottleneck. Attention mechanism. Attention outside NLP.
Transformers in NLP
Self-attention technique. Transformer architecture overview.
Contextual Embeddings
Transformer-based contextual embeddings. ELMo, BERT, GPT, XLM, etc. overview.
Large Language Models
Language modelling: recap. Compute-optimal LLMs (Chichilla scaling laws). Limitations and applicability to different tasks. Business applications.
Pretrain LLMs
How to gather data for LLM pretrain. Evaluation process and popular benchmarks. Overview of popular LLMs (ChatGPT, Gemini, Llama, Claude, etc.)
Posttrain and Alignment
Supervised fine-tune. PEFT. Adapters. LoRA, QLoRA. Prefix tuning vs P-tuning vs Prompt tuning. Prompt engineering, RAG. LLM alignment, security and privacy.
Prompt Engineering and Multimodal LLM
Prompt engineering, RAG. Overview of popular multimodal LLMs (Gemini Pro vision, SORA). Scaling laws для mixed-modal LLMs. Q&A section.
Midterm Test
NLP open problems. Discussion, section outro.
Introduction to Reinforcement Learning
Reinforcement learning problem statement. Stochastic and black-box optimization.
Value-based Methods in RL
Discounted reward in RL. Value iteration. Policy iteration.
Model-free Learning. Q-learning, SARSA
On-policy and off-policy algorithms. Approximate Q-learning. DQN. Experience replay.
Policy Gradient Methods
Policy gradient. REINFORCE algorithm.Policy gradient for sequence modelling.
Final exam
RL open problems, discussion. RL in NLP. Reinforcement learning from human feedback. Course outro.
Prerequisites
Master’s Machine Learning course (DS406) or equivalent, e.g., Introduction to Deep Learning and Computer Vision course.
Python programming experience, PyTorch basics.
At least basic knowledge of Linear Algebra, Probability Theory, Optimisation.
Methodology
The course will be organized in three-hour sessions and self-study practical assignments. Sessions will contain both theoretical and practical parts with different ratios depending on the materials.
Grading
Anastasia Ianina got her PhD from Moscow Institute of Physics and Technology where she focused on Natural Language Processing and Exploratory Search problems. She received thorough knowledge of math and machine learning, and gained significant amounts of hands-on experience: interning at Lyft and working on self-driving cars, working as a researcher at the MIPT machine intelligence lab, holding research scientist positions at Yandex, Samsung and Meta, leading teams responsible for LLM training at Tinkoff bank and WB Tech, and also writing papers to top-level international conferences.
Anastasia’s research interests include Machine Learning, Natural Language Processing, Text Analytics and Large Language Models. She currently teaches MIPT students machine learning and takes part in creating online educational courses and textbooks: she authored the course “Dynamic Neural Network Programming with PyTorch” for Packt Publishing, worked on Coursera NLP specialisation and co-authored ML textbook for Yandex School of Data Analysis.
See full profileMichael Diskin's academic and professional journey in AI and machine learning began with a BSc in Computer Science from HSE University, followed by master-level courses at the Yandex School for Data Analysis. At Wildberries, he leads the LLM R&D team, emphasising distributed training of large language models, marked by his significant contributions to NeurIPS and ICML.
His career spans roles from AI research at Rask.AI and Yandex Research, to technological innovations at Huawei and software engineering at Yandex. Michael has authored influential publications, engaged in high-level machine learning competitions, and demonstrated his commitment to education through teaching and mentoring roles at Coursera, HSE University, and YSDA.
See full profileApply for this course
Deep Learning in Applications
by Anastasia Ianina, Michael Diskin
Total hours
45 Hours
Dates
Apr 08 - Apr 26, 2024
Fee for single course
€1500
Fee for degree students
€750
How to secure your spot
Complete the form below to kickstart your application
Schedule your Harbour.Space interview
If successful, get ready to join us on campus
FAQ
Will I receive a certificate after completion?
Yes. Upon completion of the course, you will receive a certificate signed by the director of the program your course belonged to.
Do I need a visa?
This depends on your case. Please check with the Spanish or Thai consulate in your country of residence about visa requirements. We will do our part to provide you with the necessary documents, such as the Certificate of Enrollment.
Can I get a discount?
Yes. The easiest way to enroll in a course at a discounted price is to register for multiple courses. Registering for multiple courses will reduce the cost per individual course. Please ask the Admissions Office for more information about the other kinds of discounts we offer and what you can do to receive one.