November 11th
Day 2. Science and AI Technologies
Timing
Section
Streams
10:00-10:15
Day Opening
Main stage

Alexander Vedyakhin
First Deputy Chairman of the Executive Board, Sberbank

Dmitry Chernyshenko
Deputy Prime Minister of the Russian Federation
10:15-10:45
AI Cloud & Beyond
Main stage

David Rafalovsky
Sber

Greg Lavender
Intel
10:45-11:15
AIRI Presentation
Main stage

Maxim Yeremenko
Sber

Leonid Zhukov
AIRI (Artificial Intelligence Research Institute)
11:15-13:15
Keynote Talks
Main stage
Geometric deep learning: from Euclid to drug design

Michael Bronstein
Imperial College London, Twitter
Complex systems for AI

Tomas Mikolov
CIIRC CTU
The most popular artificial neural networks

Juergen Schmidhuber
IDSIA, AIRI (Artificial Intelligence Research Institute)
Automatic visual recognition: from pixels to machines that see, reason and act

Josef Sivic
CIIRC CTU, ELLIS Unit Prague
11:15-12:15
Expert Talks
AI research stream
Building a server for AI deploy

Mikhail Tsvetkov
Intel
AI based musical style-transfer

Ivan Shanin
Huawei
12:15-13:15
Discussion: «AI for Science, Science for AI»
AI research stream
AI for Science, Science for AI

Albert Yefimov
Sber

Konstantin Novoselov
Russian and British physicist. Nobel Prize laureate, Fellow of the Royal Society of London, International Member of the National Academy of Sciences, Co-chairman of the Sber Scientific Prize Committee

Stanislav Smirnov
Russian and Swiss mathematician, Fields Medal winner. Academic supervisor at the Department of Mathematics and CS of SPbU, Prof. at the University of Geneva and Skoltech, Director of the NCCR SwissMAP. Co-founder of the Sirius Educational Center

Konstantin Anokhin
Russian neurobiologist, Academician of the Russian Academy of Sciences. Director of the Institute for Advanced Brain Research at MSU, Head of the Laboratory for Neurobiology of Memory at P. K. Anokhin Research Institute of Normal Physiology
13:15-13:40
Break
13:40-18:40
Keynote Talks
Main stage
Scalable and Sustainable AI Acceleration for Everyone: Algorithms Train Billion-parameter AI Models on a Commodity CPU faster than Hardware Accelerators

Anshumali Shrivastava
Rice University
Siri, Let Me Introduce Siri: The Connected Future of Artificial Intelligence

Michael Wooldridge
University of Oxford
Deep Problems of Deep Learning

Mikhail Burtsev
AIRI (Artificial Intelligence Research Institute), MIPT
Reinforcement Learning for Real-World System Optimization

Somdeb Majumdar
Intel
Motion Prediction in Robotic Tasks

Gonzalo Ferrer
Skoltech
The Robustness of Deep Networks: A Geometrical Perspective

Evgeny Burnaev
Skoltech, AIRI (Artificial Intelligence Research Institute)
Geometry in Machine learning

Ivan Oseledets
Skoltech
Amazing qualities of the loss function relief in deep learning

Dmitry Vetrov
HSE University, AIRI (Artificial Intelligence Research Institute)
Deploy and Run your Cloud Native AI Apps Anywhere

Mo Haghighi
IBM
13:40-17:00
Sber Session
AI research stream

Dmitrii Krainov
Sber
Multimodal neural networks ruDALLE + ruCLIP

Sergey Markov
Sber

Denis Dimitrov
Sber
Confidential computing & data preparation for AI models’ training

Boris Rabinovich
Sber
AI/ML product delivery. Easy road

Otari Melikishvili
SberCloud
Graphs as the way to boost your business though ML and otherwise

Ilya Pozdnyakov
Sber

Mikhail Stepnov
Sber
Digitize everything. SberOCR

Marat Gazizov
Sber

Stepan Evstifeev
Sber
Adversarial attacks on transactional models

Ivan Kireev
Sber
The conversationalist: what lies behind chitchat in Salute virtual assistants family

Denis Kirjanov
Sber
Contrastive learning on event sequence data

Dmitrii Babaev
Sber, AIRI (Artificial Intelligence Research Institute)
Between interpretability and performance: Why we need GlassBox models

Oleg Travkin
Sber
Python Library SberProcessMining(SberPM)

Andrey Bugaenko
Sber
13:40-18:40
Workshops
Workshops
Mindspore platform for training neural network models

Dmitry Yudin
AIRI (Artificial Intelligence Research Institute), MIPT

Andrey Kochetov
Huawei
Training computer vision models in the Mindspore framework

Vitaliy Bezugliy
MIPT

Andrey Kochetov
Huawei
Code generation technologies today: will we replace the programmers or solve the development problems?

Tatiana Shavrina
Sber, AIRI (Artificial Intelligence Research Institute)

Andrey Kochetov
Huawei
Neuromorphic computing in fundamental research and application development of future magnetometry

Alexandra Bernadotte
Sber

Yulia Sandamirskaya
Intel
Self-supervised learning for sequential data

Alexey Zaytsev
Skoltech
Handwritten Text Recognition

Mark Potanin
Sber

Aleksandr Korotaevskiy
Sber
DeepPavlov in 2021: From NLP Tasks to Dev Tools for Conversational AI

Daniel Kornev
DeepPavlov (MIPT)
DialogFlow Framework: create a smart bot just in a few steps

Denis Kuznetsov
DeepPavlov (MIPT)
Neural-symbolic integration: Scene Representation and Multi-agent Reinforcement Learning

Aleksandr Panov
AIRI (Artificial Intelligence Research Institute), MIPT

Alexey Skrynnik
AIRI (Artificial Intelligence Research Institute)
Learn Together, Stop Apart: a Novel Approach to Ensemble Pruning

Gleb Gusev
Sber, AIRI (Artificial Intelligence Research Institute)

Bulat Ibragimov
AIRI (Artificial Intelligence Research Institute), MIPT
Driving Business Success with Crowdsourcing: Practical Use Cases

Roman Gaev
Yandex
RecSys RePlay Workshop

Alexey Vasilev
Sber

Yan-Martin Tamm
Sber
13:40-15:00
AGI
Technology sections

Anton Kolonin
Aigents, SingularityNET
Challenges of training large-scale neural networks as a significant barrier on the way to Artificial General Intelligence

Vitaly Milke
Bauman MSTU
OOD detection

Maxim Panov
Skoltech
The Future of dialogue systems: new implementations of ruGPT-3

Valery Ternovsky
Sber
Emergent intelligence and distributed stochastic optimization

Konstantin Amelin
Saint Petersburg University
13:40-15:20
NLP
Technology sections

Loran Jacobs
iPavlov
From Alexa Prize Challenge to DeepPavlov DREAM

Dilyara Baymurzina
DeepPavlov (MIPT)
Cross-Lingual Question Answering as Information Retrieval and Text Generation Task

Valentin Malykh
Huawei Moscow Research Center
Low- and No-code instruments for NLP tasks

Prokhor Gladkikh
Sber

Aleksander Korekov
Sber
Beyond Shallow Embeddings of Knowledge Graphs

Michael Galkin
Mila Quebec AI Institute & McGill University
Neural Entity Linking: A Survey of Models Based on Deep Learning

Alexander Panchenko
Skoltech
13:40-15:20
Intel: AI tools for developers
Technology sections

Dmitry Sivkov
Intel
Accelerating inference for fast simulation at CERN with oneAPI

Sofia Vallecorsa
CERN
Intel oneAPI tools in SberCloud ML Space

Vladimir Novozhenov
SberCloud
Project Cable Checker & oneAPI tools

Mikhail Elin
Digital Wave
OpenVINO 2.0 — new NLP opportunities

Alexander Kozlov
Intel

Vasily Shamporov
Intel
15:00-17:00
CV
Technology sections

Dmitry Markov
VisionLabs
Face Recognition Today: Recent Advances and Open Challenges

Sergey Miliaev
VisionLabs
Video analytics for monitoring security in public areas from requirements collection to realization

Vladimir Kravcev
Sber

Alexey Belikov
Sber

Tatyana Voronova
Sber
Image Neural Editing 2021

Roman Suvorov
Samsung AI Center
ISyNet: Design of Pareto-optimal Neural Architecture for Computer Vision tasks on Neural Processor Family

Vladimir Korviakov
Huawei
3D Vision - depth estimation, depth completion and 3D detection

Anton Konushin
Samsung AI Center Moscow, Lomonosov Moscow State University
Neural Network Video Analysis of People's Movements

Daniel Kireev
VisionLabs
15:20-17:05
RecSys
Technology sections

Ivan Kuzmin
Sber
How to get bandits to the bank?

Dmitry Bugaychenko
Sber
Mimicking attention mechanisms for fast and accurate sequential recommendations

Evgeny Frolov
Skoltech
Improving Long-Term Functioning of a Recommendation System using RL

Alexander Fritzler
Yandex
OpenRecSys benchmark

Alexey Vasilev
Sber

Alexey Natekin
Data Souls, Open Data Science
Optimization of machine learning and graph analysis algorithms on processors with ARM architecture for Apache Spark clusters

Stanislav Moiseev
Huawei
15:20-18:55
AI in Russia
Technology sections
Langrange сharacteristics application in Pacific Ocean open water for short-term catch location prediction by airborne traps

Vladimir Kulik
Pacific Branch of the Russian Federal Research Institute of Fisheries and Oceanography
Machine learning for pelagic fishing support in the Northwest Pacific

Igor Chernienko
Far Eastern Branch, Russian Academy of Sciences, V.I. Il'ichev Pacific Oceanological Institute
Why is it important to explain a loan decline to the client?

Igor Barkhatov
Sber
Automatic recognition of objects in images using the DLgram cloud service

Andrey Matveev
Novosibirsk State University
AI: at the intersection of science and technology

Tatiana Zobnina
Naumen
Neuromorphic intelligent systems based on models of spike neural networks of the brain

Victor Kazantsev
IKBFU, Lobachevsky University (UNN), Innopolis University
Explainable AI: How to gain trust in Machine Learning-based solutions

Andrey Derevyanka
EPAM Systems
Photonics for smart home and smart city

Victor Soifer
Samara National Research University
How to open the AI black box?

Tatiana Gavrilova
Saint Petersburg University
AI-based Optimal Control for Energy Storage Systems

Ovanes Petrosian
Saint Petersburg University
MDMMT: Multi-domain Multi-modal Transformer — SotA Solution for Video Retrieval Task

Maksim Dzabraev
Huawei
Workout control

Petr Vol'nov
ITMO University
Does AI dream of evolution?

Anna Kalyuzhnaya
ITMO University
AI in Energy

Maxim Shcherbakov
Volgograd State Technical University (VSTU)
Intelligent Patent Processing

Alla Kravets
Volgograd State Technical University (VSTU)
17:00-18:40
AIRI Session
AI research stream
AI for Materials Science & Sustainable Energy

Semen Budennyy
AIRI (Artificial Intelligence Research Institute)
Computer-aided approaches for vaccine design

Nikita Ivanisenko
AIRI (Artificial Intelligence Research Institute)
Using AI to interpret non-coding variants in the human genome

Veniamin Fishman
AIRI (Artificial Intelligence Research Institute)
Ab initio AI

Arthur Kadurin
AIRI (Artificial Intelligence Research Institute)
Efficient Deep Active Learning for Text Processing. The New Old Way to Deal with Annotated Data Scarcity

Artem Shelmanov
AIRI (Artificial Intelligence Research Institute)

Akim Tsvigun
Sber
17:00-18:40
AutoML
Technology sections
Open source AutoML benchmark

Alexander Ryzhkov
Sber

Alexey Natekin
Data Souls, Open Data Science
Pyboost: Research tool to explore gradient boosting

Anton Vakhrushev
Sber

Leonid Iosipoi
AIRI (Artificial Intelligence Research Institute)
AutoML and deep learning for tabular data processing

Ivan Bondarenko
Huawei Novosibirsk Research Center
FEDOT - open-source evolutionary AutoML for composite pipelines

Nikolay Nikitin
ITMO University
17:05-18:40
Speech Analytics
Technology sections

Dmitry Dyrmovsky
Speech Technology Center
Voice biometrics in uncontrolled conditions

Alexey Sholokhov
Huawei
Out of call center: new challenges for speech analytics

Oleg Petrov
Speech Technology Center
Voice Scoring. What can we get from dialogues with the client?

Alexander Demidov
ActiveBusinessConsult
AI in telemarketing or a Robot that knows what you need

Elena Larina
Sber
Using unlabeled data in speech recognition tasks

Nikita Semenov
MTS AI
18:40-19:55
AIIJC Awards
Main stage
19:55-20:00
Day Closure
Main stage