OpenTalks.AI /
19-20 Февраля 2026
Белград, Сербия
Открытая конференция по ии
These videos are intended exclusively for participants of the OpenTalks.AI conference.
Distribution of video recordings will be a violation of the agreement to participate in the conference, which the participant agreed to upon registration.
The rest of the videos and slides will be added later.
Distribution of video recordings will be a violation of the agreement to participate in the conference, which the participant agreed to upon registration.
The rest of the videos and slides will be added later.
Neuromorphic computing
Neurosemantic Network: An Alternative to LLMs Based on Spiking Neural Networks - Andrey Lavrentyev, Kaspersky Lab
Neuromorphic AI: Towards Practical Implementation in Diverse Application Domains - Oleg Vygolov, Kaspersky Lab
Deep Convolutional Spiking Neural Networks - Mikhail Kiselev, Kaspersky Lab
Neurosemantic Network: An Alternative to LLMs Based on Spiking Neural Networks - Andrey Lavrentyev, Kaspersky Lab
Neuromorphic AI: Towards Practical Implementation in Diverse Application Domains - Oleg Vygolov, Kaspersky Lab
Deep Convolutional Spiking Neural Networks - Mikhail Kiselev, Kaspersky Lab
LLM - R&D
Overview: LLM Pre-training in 2025 - Vladislav Savinov, Yandex
LLMs in Conversational Agents for Mental Health Care - Overview of
Recent Research, Key Challenges and Future Directions - Alexander Kotov, Wayne State University
Overview: LLM Pre-training in 2025 - Vladislav Savinov, Yandex
LLMs in Conversational Agents for Mental Health Care - Overview of
Recent Research, Key Challenges and Future Directions - Alexander Kotov, Wayne State University
LLM & Graph & RAG
Explainable Document Classification via Concept Whitening and Stable Graph Patterns - Sergey Kuznetsov, HSE University
GraphRAG Meets Cyrillic: Adapting Graph-Retrieval-Augmented Generation for Russian - Mariya Godunova, HSE University, Alina Avanesyan, HSE University
Explainable Document Classification via Concept Whitening and Stable Graph Patterns - Sergey Kuznetsov, HSE University
GraphRAG Meets Cyrillic: Adapting Graph-Retrieval-Augmented Generation for Russian - Mariya Godunova, HSE University, Alina Avanesyan, HSE University
LLMs in Business
Programmatic Approach to Human Intelligence for AI: Toloka Platform - Sergey Polyashov, Toloka
Using LLM to Improve Incident Handling in Azure - Alexander Svetkin, Microsoft
Scaling integrations at Incode using spec-driven development - Predrag Radenkovic, Codeplain
AI Is Your New Colleague: A Task-Centric Recommendation System for Interpretable Work Augmentation - Chiara Stramaccioni, Karimi
Programmatic Approach to Human Intelligence for AI: Toloka Platform - Sergey Polyashov, Toloka
Using LLM to Improve Incident Handling in Azure - Alexander Svetkin, Microsoft
Scaling integrations at Incode using spec-driven development - Predrag Radenkovic, Codeplain
AI Is Your New Colleague: A Task-Centric Recommendation System for Interpretable Work Augmentation - Chiara Stramaccioni, Karimi
Overviews
Top computing systems for HPC and AI - Valery Yegorshev, Cognitar LLC
Top computing systems for HPC and AI - Valery Yegorshev, Cognitar LLC
CV in business
Computer Vision for Ore Pass Functioning Control - Aleksandr Rassadin, Severstal
One GPU, Hundred Eyes: Real-Time Multi-Camera Analytics for Cargo-Drop Detection on the Edge - Mikhail Krasilnikov,
Bia-technologies
Finding a Middle Ground Between Industrial Automation and Robotics for Effective Business Solutions - Egor Ershov,
MIPT, AIRI
Deploying Deepfake Detection in a Production VCS: The Kontur.Talk Case - Pavel Kuznetsov,
Kontur
Computer Vision for Ore Pass Functioning Control - Aleksandr Rassadin, Severstal
One GPU, Hundred Eyes: Real-Time Multi-Camera Analytics for Cargo-Drop Detection on the Edge - Mikhail Krasilnikov,
Bia-technologies
Finding a Middle Ground Between Industrial Automation and Robotics for Effective Business Solutions - Egor Ershov,
MIPT, AIRI
Deploying Deepfake Detection in a Production VCS: The Kontur.Talk Case - Pavel Kuznetsov,
Kontur
CV - Research
How do I find everything? Let's scale from 1 to 100 diseases - Evgenii Nikitin,
Celsus AI
Deep Learning-Based Multi-Object Tracking for Nonlinear Motion - Momir
Adžemović, University of Belgrade
How do I find everything? Let's scale from 1 to 100 diseases - Evgenii Nikitin,
Celsus AI
Deep Learning-Based Multi-Object Tracking for Nonlinear Motion - Momir
Adžemović, University of Belgrade
Generative AI - research
How to create a viral AI sticker pack generator based on users’ photos - Natalia Khanzhina,
Independent Researcher
Generative AI–based solutions at Lemana PRO (ex-Leroy Merlin) - Ksenija Blažević,
Lemana Tech
How to create a viral AI sticker pack generator based on users’ photos - Natalia Khanzhina,
Independent Researcher
Generative AI–based solutions at Lemana PRO (ex-Leroy Merlin) - Ksenija Blažević,
Lemana Tech
GenAI - academic track
RusCode: Russian Cultural Code Benchmark for Text-to-Image Generation - Julia Agafonova, Kandinsky Lab
A Brief History of Visual Generative AI: Kandinsky models - Viacheslav Vasilev,
Kandinsky Lab
RusCode: Russian Cultural Code Benchmark for Text-to-Image Generation - Julia Agafonova, Kandinsky Lab
A Brief History of Visual Generative AI: Kandinsky models - Viacheslav Vasilev,
Kandinsky Lab
Reinforcement learning
Zero-Sum Positional Differential Games as a Framework for Robust Reinforcement Learning: Deep Q-Learning Approach - Anton Plaksin,
Nebius
HGRPO: Hierarchical Grouped Reward Policy Optimization for Multi-Turn Conversational Agents - Karina Romanova, Yandex
Zero-Sum Positional Differential Games as a Framework for Robust Reinforcement Learning: Deep Q-Learning Approach - Anton Plaksin,
Nebius
HGRPO: Hierarchical Grouped Reward Policy Optimization for Multi-Turn Conversational Agents - Karina Romanova, Yandex
LLM in Business
Data mining based on large language models - Evgenii Grigorev,
T1.Artificial Intelligence
Avibe: How and Why They Did an LLM on Avito - Anastasiia Rysmiatova,
Avito
How AI Agents Replace Manual Analytics in Oil & Gas Operations - Dmitrii Pshichenko,
NIS A.D.
Dalibor Lazarevic,
NIS A.D
Data mining based on large language models - Evgenii Grigorev,
T1.Artificial Intelligence
Avibe: How and Why They Did an LLM on Avito - Anastasiia Rysmiatova,
Avito
How AI Agents Replace Manual Analytics in Oil & Gas Operations - Dmitrii Pshichenko,
NIS A.D.
Dalibor Lazarevic,
NIS A.D
LLM development
From Projects to Product: How We Systematized Our Work with LLM -
Vladislav Balaev,
Lanit-technology
LLM tool calling and context management - Dmitry Vasyuk,
Microsoft
LLM-Assisted Grading of Open-Ended Student Responses - Jelena Graovac,
University of Belgrade
From Projects to Product: How We Systematized Our Work with LLM -
Vladislav Balaev,
Lanit-technology
LLM tool calling and context management - Dmitry Vasyuk,
Microsoft
LLM-Assisted Grading of Open-Ended Student Responses - Jelena Graovac,
University of Belgrade
LLM research
Production-Ready Adapters for On-Device Language Models - Marat Saidov,
Microsoft
Production-Ready Adapters for On-Device Language Models - Marat Saidov,
Microsoft
Neurocognitive architectures
Cognitive Architecture for Neuro-Symbolic Experiential Learning - Anton Kolonin,
Aigents
Implementing Digital Self-Awareness Based on a Cognitive-Neuromorphic Approach - Alexei Samsonovich,
George Mason University, NRNU MEPhI
An elementary universal cycle of continuous attention as a new model for computing meta-attractor generalizations - Aleksey Kabanov,
BTR R&D
Cognitive Architecture for Neuro-Symbolic Experiential Learning - Anton Kolonin,
Aigents
Implementing Digital Self-Awareness Based on a Cognitive-Neuromorphic Approach - Alexei Samsonovich,
George Mason University, NRNU MEPhI
An elementary universal cycle of continuous attention as a new model for computing meta-attractor generalizations - Aleksey Kabanov,
BTR R&D