The program of the AI Autumn School includes five mini-courses on machine learning utilizing innovative algorithms such as Stochastic Gradient Descent (SGD), convex online optimization and its unexpected applications, transformers in computer vision, and tabular deep learning. The courses will be delivered by leading scholars in the field of AI: Martin Takac (Mohammed bin Zayed University of Artificial Intelligence), Francesco Orabona (King Abdullah II School of Information Technology), Artem Babenko (Yandex), Sergey Zagoruyko (MTS AI Center), and Vladimir Spokoiny (Weierstrass Institute, Humboldt University of Berlin, HSE).

Registration for listeners is open until October 5. The number of available places is limited.

The organizers of ASCOMP 2024 are Innopolis University, Artificial Intelligence Institute at Moscow State University, and the Steklov Mathematical Institute of the Russian Academy of Sciences.

ASCOMP Program:

7 October

⦁ 10:00 — 11:00  

  Sergey Zagoruyko: Transformers in Computer Vision. Transformer Architecture

⦁ 11:10 — 12:10  

  Sergey Zagoruyko: Transformers in Computer Vision. Transformers in Object Detection

⦁ 12:20 — 13:20  

  Artem Babenko: Tabular Deep Learning. Lecture 1

⦁ 15:00 — 16:00  

  Vladimir Spokoiny: Linearly Perturbed Optimization with Applications to Machine Learning. Lecture 1

⦁ 16:20 — 17:20  

  Martin Takac: Unlocking Machine Learning Potentials with Innovative SGD Algorithms. Lecture 1

⦁ 17:30 — 18:30  

  Martin Takac: Unlocking Machine Learning Potentials with Innovative SGD Algorithms. Lecture 2

8 October

⦁ 10:00 — 11:00  

  Sergey Zagoruyko: Transformers in Computer Vision. Transformers in Autonomous Driving

⦁ 11:10 — 12:10  

  Artem Babenko: Tabular Deep Learning. Lecture 2

⦁ 12:20 — 13:20  

  Martin Takac: Unlocking Machine Learning Potentials with Innovative SGD Algorithms. Lecture 3

⦁ 15:00 — 16:00  

  Vladimir Spokoiny: Linearly Perturbed Optimization with Applications to Machine Learning. Lecture 2

⦁ 16:20 — 18:00  

  Poster Session

9 October

⦁ 10:00 — 11:00  

  Francesco Orabona: Online Convex Optimization and Its Surprising Applications. Lecture 1

⦁ 11:10 — 12:10  

  Francesco Orabona: Online Convex Optimization and Its Surprising Applications. Lecture 2

⦁ 12:20 — 13:00  

  Francesco Orabona: Online Convex Optimization and Its Surprising Applications. Lecture 3

Working language of the school: English.

————————

ICOMP 2024, the First International Conference on Computational Optimization, will take place on October 10–12 in Innopolis city. At this conference, global experts in the fields of artificial intelligence and mathematics will discuss machine learning methods, edge computing, and other issues related to numerical optimization methods, including production process planning, logistics, engineering design, telecommunications, cloud technologies, bioinformatics, and nanotechnology.