[ИАД, весна 2026] Рекомендательные системы. Лекция 1

Value of the Information & Strength of the Argument

The lecture provides a solid foundation in recommender systems, covering essential concepts and metrics. The argumentation is clear and logical, building from historical context to formal problem definition and evaluation. The instructor effectively explains the rationale behind different metrics and the importance of rigorous experimental design. The content is valuable for students and practitioners seeking a structured introduction, though it does not delve into advanced techniques or recent developments in deep learning-based recommenders.

Scientific Rigor, Source Quality, Title Accuracy

The lecture demonstrates scientific rigor by referencing classic textbooks and seminal papers, such as those by Adomavicius and Tuzhilin, and by discussing evaluation pitfalls. The title accurately reflects the content. The instructor’s expertise is evident, and the presentation is well-organized. However, the lecture does not provide detailed citations for all claims, and the references are not exhaustive. The adéquation between title and content is strong, as it is indeed the first lecture of a course on recommender systems.

The title accurately reflects the content: it is the first lecture of a course on recommender systems, providing an introduction and overview.

8/10

The lecture is a structured academic introduction to recommender systems, covering historical context, problem formulation, metrics, and evaluation pitfalls. The content is technically accurate and aligns with established knowledge in the field. The instructor demonstrates expertise and provides references to classic literature. However, as a single lecture, it lacks the depth of a comprehensive review, and some claims (e.g., about Netflix Prize) are presented without detailed evidence.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 00:00Introduction and waiting for participants
• 02:00Start of lecture: historical definitions of recommender systems
• 05:00Motivation: business and user benefits
• 10:00Overview of recommender system types
• 15:00Formal problem formulation: users, items, feedback matrix
• 20:00Evaluation metrics: NDCG and MAP
• 30:00Experimental design and evaluation pitfalls
• 40:00Course logistics and homework information

• Recommender Systems Handbook — Mentioned as classic literature for recommender systems.
• Recommender Systems: An Introduction — Mentioned as classic literature.
• Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions — Referenced as a paper introducing a modern problem formulation.
• Evaluating Recommender Systems — Referenced as an example of a survey on evaluation.

• Recommender Systems Handbook — Classic reference that aligns with the lecture's content on recommender systems.
• Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions — The lecture's problem formulation aligns with this seminal paper.