2026/2027





Рекомендательные системы
ID 1121366
Статус:
Маго-лего
Где читается:
Факультет компьютерных наук
Когда читается:
1, 2 модуль
Охват аудитории:
для всех кампусов НИУ ВШЭ
Преподаватели:
Джин Сеунгмин
Язык:
английский
Кредиты:
6
Контактные часы:
40
Course Syllabus
Abstract
In this course, we will introduce the problem of building a recommender system (RS) and its relation to other domains of machine learning and information retrieval. We will start by providing an overview of classical approaches for constructing RSs, including content-based and collaborative filtering via matrix factorization. Additionally, we will discuss the metrics and validation schemes commonly employed in RS development.Moving forward, we will delve into modern neural architectures specifically designed for recommender systems. Furthermore, we will explore various techniques frequently utilized in the industry, such as session-based recommender systems, two-stage RSs, and online RSs. Lastly, we may touch upon additional topics of common interest.
Learning Objectives
- Introduction to classical and modern models and methods of recommender systems . Expanding the practical skills of a data science specialist .
Expected Learning Outcomes
- Creation of a prototype of a recommender system based on collaborative filtering methods
- Finding patterns in data ( association rules , frequent sets of items and subsequences of events ).
- Building a taste profile of the user and products
- Understanding relevant quality measures in the field of recommender systems
- Conducting practical research in the field of recommender systems
Course Contents
- Introduction to recommender systems . Taxonomy of recommender systems . Case-study examples . Methods for assessing the quality of recommender systems .
- Content-Based Filtering
- Collaborative filtering methods . Case-study: User-based and item-based approaches . Bimodal cross-validation . Movie Lens Dataset.
- Frequent sets of goods . Association rules . Case-study: Contextual Advertising.
- Matrix factorization methods . Case-study: Boolean matrix factorization (BMF), non-negative matrix factorization (NMF), singular value decomposition (SVD).
- Social Recommender with PageRank
- Deep Learning Recommender Systems
- Designing LLM4RecSys: Architecting End-to-End Recommendation Pipelines via Encoders, Rankers, RAG, and Routing
- Capstone Project
Assessment Elements
- Code ReviewStudents explain their code in their own words to show their level of understanding.
- Capstone Project . Small Groups
- Midterm Exam: HackathonStudents developing a recommendation system for a Kaggle competition during a 24-hour hackathon.
- Class Activity
- Attendance
Interim Assessment
- 2026/2027 2nd module**Final Mark** The final score is calculated on a **100-point scale** based on the following components: * attendance: 10%; * homework: 20%; * class activity: 10%; * 24-hour Kaggle midterm competition: 20%; * final capstone project: 40%. Homework assignments include practical notebooks, brief interpretations of results, and other individual assignments. Class activity includes practical tasks completed during class, discussion of results, code review, and peer feedback. Students must be able to explain the baseline, the changes made, their impact on the metric, and the limitations of an experiment. **Determination of the Final Mark** Each student's final score is calculated out of 100 points. To obtain the **base mark on the 10-point scale**, the final score is divided by 10 and the result is rounded to the nearest whole number. For example: * 68 points → 6.8 → **7 points**; * 84 points → 8.4 → **8 points**; * 85 points → 8.5 → **9 points**; * 96 points → 9.6 → **10 points**. Students who have met all mandatory course requirements are then ranked according to their final scores. **Ranking-based Cut-off Rule** The ranking-based cut-off is applied only when the number of students receiving a particular mark exceeds the allowed proportion of the cohort. If more than **30% of students receive a mark of 8 or higher**, only students within the top 30% of the ranking retain their converted marks of 8, 9, or 10. Students outside the top 30% whose converted mark is 8 are assigned **7**. If more than **70% of students receive a mark of 6 or higher**, students outside the top 70% whose converted mark is 6 are assigned **5**. The boundaries between the groups may be adjusted by no more than ±5 percentage points in light of the actual distribution of final scores. **Blocking Requirements** The final capstone project is a mandatory **blocking assessment component**. Failure to submit the capstone project results in an **unsatisfactory final mark**, regardless of the student's accumulated score in other components. A student who obtains **fewer than 35 points out of 100** also receives an **unsatisfactory final mark**, regardless of their position in the ranking. For students who submit the capstone project and obtain at least 35 points, the final mark on the **10-point scale** is determined based on the base mark and the ranking-based cut-off rule described above.
Bibliography
Recommended Core Bibliography
- Integrating deep learning algorithms to overcome challenges in big data analytics, , 2022
- Manouselis, N., Drachsler, H., Verbert, K., Duval, E. Recommender Systems for Learning. – Springer, 2013. – ЭБС Books 24x7.
- Pro Deep Learning with TensorFlow 2.0 : a mathematical approach to advanced artificial intelligence in Python, Pattanayak, S., 2023
- Time series algorithms recipes : implement machine learning and deep learning techniques with Python, , 2023
- Transformers for machine learning : a deep dive, Kamath, U., 2022
Recommended Additional Bibliography
- Parul Aggarwal, Vishal Tomar, & Aditya Kathuria. (2017). Comparing Content Based and Collaborative Filtering in Recommender Systems. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsbas&AN=edsbas.32D5064E