Full Transcript
Loading transcript…
AI: от основ до трансформеров. Об этом курсе и краткая история AI. Лекция 1.
Watch with subtitles, summary & AI chat
Add the free Subkun extension — works directly on YouTube.
- Watch
- Subtitles
- Summary
- Ask AI
Students and developers seeking a structured, updated academic path from classical machine learning basics to advanced deep learning and transformer models.
TL;DR
This introductory lecture launches a new, consolidated course on machine learning and deep learning. The instructor explains the curriculum shift driven by the rapid evolution of large language models and outlines the transition from foundational algorithms to modern transformer architectures.
Key Takeaways
- The rapid evolution of artificial intelligence, particularly the rise of large language models, necessitates a fundamental restructuring of educational curricula to prioritize modern deep learning techniques over traditional algorithms.
- Traditional machine learning algorithms like decision trees and gradient boosting, while mathematically elegant, risk becoming obsolete in academic settings due to the overwhelming dominance of contemporary neural network architectures.
- Developing complex, specialized AI systems for niche use cases remains a challenging but essential business model, as these problems often cannot be solved by off-the-shelf models found on platforms like GitHub.
- Defining artificial intelligence involves a tension between mimicking human cognitive processes and pursuing purely rational, goal-oriented behavior, with the latter often proving more effective for practical problem-solving.
- The historical cycle of AI hype, where startups frequently leverage the term to attract funding, mirrors previous technological booms, suggesting a need for critical evaluation of AI's actual utility versus its marketing.
- Effective AI education requires balancing foundational mathematical theory with hands-on experience in cutting-edge fields like computer vision and natural language processing to ensure students remain at the industry's forefront.
In This Video
- 00:00Введение и структура курса
Евгений Резников представляет обновленный курс, объединяющий машинное и глубокое обучение в ответ на стремительное развитие технологий.
- 02:16Опыт и профессиональный путь
Лектор рассказывает о своем восьмилетнем опыте в AI, работе в компаниях FIX и Prevision Technologies, а также о своих образовательных проектах.
- 05:23Обзор учебных материалов
Обсуждается дерево курсов автора, включая математику для машинного обучения, компьютерное зрение и работу с трансформерами.
- 07:31Ответы на вопросы студентов
Евгений отвечает на организационные вопросы аудитории касательно формата лекций, доступности записей и качества материалов.
- 08:15Определение искусственного интеллекта
Лектор анализирует различные подходы к пониманию AI, сравнивая человеческое мышление, рациональное мышление и практические действия систем.
Questions & Answers
What is the structure of this new AI course?
The course combines traditional machine learning and deep learning into one curriculum. It covers basic machine learning in the first 5-6 lectures, which then serves as a foundation for advanced topics like computer vision, natural language processing, and anomaly detection.
Why did the instructor merge machine learning and deep learning courses?
Two main reasons: the rapid evolution of AI, specifically large language models, makes traditional non-deep learning algorithms feel less relevant, and the instructor now has two lectures per week, allowing for a more integrated and efficient teaching approach.
What is the instructor's professional background in AI?
Evgeny Reznikov has been lecturing on AI and machine learning for 8 years. He leads the machine learning and computer vision department at FIX and co-founded Prevision Technologies, a company that develops custom AI systems for complex cases not covered by standard open-source models.
What is the 'Centi' project mentioned in the lecture?
Centi is a project founded by the instructor that uses machine learning and artificial intelligence to help people understand and navigate the world of perfumery, currently available as a Telegram bot with mobile applications in development.
How does the instructor define artificial intelligence?
The instructor discusses AI through four lenses: systems that think like humans, systems that think rationally, systems that act like humans, and systems that act rationally, ultimately emphasizing the importance of rational action to solve specific tasks effectively.
Key Terms
- Large Language Models — Advanced AI models that have significantly influenced the field, making traditional machine learning algorithms feel less central to modern AI development.
- Computer Vision — A field of artificial intelligence that enables computers and systems to derive meaningful information from digital images, videos, and other visual inputs.
- Gradient Boosting — A machine learning technique for regression and classification problems that produces a prediction model in the form of an ensemble of weak prediction models.
- Transformers — A type of neural network architecture that has become a standard in modern AI, particularly for tasks involving natural language processing.