Mikhail Burtsev

ORCID: 0000-0003-1614-1695
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Topic Modeling
  • Natural Language Processing Techniques
  • Speech and dialogue systems
  • Evolutionary Game Theory and Cooperation
  • Reinforcement Learning in Robotics
  • Neural Networks and Applications
  • Neural dynamics and brain function
  • Multimodal Machine Learning Applications
  • Evolutionary Algorithms and Applications
  • Evolution and Genetic Dynamics
  • Advanced Text Analysis Techniques
  • Shoulder Injury and Treatment
  • Shoulder and Clavicle Injuries
  • Bone fractures and treatments
  • AI in Service Interactions
  • Neuroscience and Neural Engineering
  • EEG and Brain-Computer Interfaces
  • Genomics and Phylogenetic Studies
  • Text and Document Classification Technologies
  • Semantic Web and Ontologies
  • Advanced Memory and Neural Computing
  • Advanced Graph Neural Networks
  • Connective tissue disorders research
  • RNA and protein synthesis mechanisms
  • Orthopedic Surgery and Rehabilitation

London Institute for Mathematical Sciences
2023-2025

Royal Institution of Great Britain
2024-2025

Albemarle (United States)
2025

Laban/Bartenieff Institute of Movement Studies
2024

Tokyo Institute of Technology
2023

Administration for Community Living
2023

IT University of Copenhagen
2023

American Jewish Committee
2023

RIKEN Center for Advanced Intelligence Project
2023

Mongolia International University
2023

Abstract Recent advancements in genomics, propelled by artificial intelligence, have unlocked unprecedented capabilities interpreting genomic sequences, mitigating the need for exhaustive experimental analysis of complex, intertwined molecular processes inherent DNA function. A significant challenge, however, resides accurately decoding which inherently involves comprehending rich contextual information dispersed across thousands nucleotides. To address this need, we introduce GENA language...

10.1093/nar/gkae1310 article EN cc-by-nc Nucleic Acids Research 2025-01-11

Mikhail Burtsev, Alexander Seliverstov, Rafael Airapetyan, Arkhipov, Dilyara Baymurzina, Nickolay Bushkov, Olga Gureenkova, Taras Khakhulin, Yuri Kuratov, Denis Kuznetsov, Alexey Litinsky, Varvara Logacheva, Lymar, Valentin Malykh, Maxim Petrov, Vadim Polulyakh, Leonid Pugachev, Sorokin, Maria Vikhreva, Marat Zaynutdinov. Proceedings of ACL 2018, System Demonstrations. 2018.

10.18653/v1/p18-4021 article EN cc-by 2018-01-01

Recent advancements in genomics, propelled by artificial intelligence, have unlocked unprecedented capabilities interpreting genomic sequences, mitigating the need for exhaustive experimental analysis of complex, intertwined molecular processes inherent DNA function. A significant challenge, however, resides accurately decoding which inherently involves comprehending rich contextual information dispersed across thousands nucleotides. To address this need, we introduce GENA-LM, a suite...

10.1101/2023.06.12.544594 preprint EN cc-by-nc-nd bioRxiv (Cold Spring Harbor Laboratory) 2023-06-13

Enabling open-domain dialogue systems to ask clarifying questions when appropriate is an important direction for improving the quality of system response. Namely, cases a user request not specific enough conversation provide answer right away, it desirable question increase chances retrieving satisfying answer. To address problem 'asking in dialogues': (1) we collect and release new dataset focused on single- multi-turn conversations, (2) benchmark several state-of-the-art neural baselines,...

10.18653/v1/2021.emnlp-main.367 article EN cc-by Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing 2021-01-01

A major limitation for the broader scope of problems solvable by transformers is quadratic scaling computational complexity with input size. In this study, we investigate recurrent memory augmentation pre-trained transformer models to extend context length while linearly compute. Our approach demonstrates capability store information in sequences up an unprecedented two million tokens maintaining high retrieval accuracy. Experiments language modeling tasks show perplexity improvement as...

10.48550/arxiv.2304.11062 preprint EN other-oa arXiv (Cornell University) 2023-01-01

Genetically encoded calcium indicators (GECIs) are mainly represented by two- or one-fluorophore-based sensors. One type of two-fluorophore-based sensor, carrying Opsanus troponin C (TnC) as the Ca2+-binding moiety, has two binding sites for ions, providing a linear response to ions. One-fluorophore-based sensors have four but better suited in vivo experiments. Herein, we describe novel design GECI with sites. The engineered called NTnC, uses TnC inserted mNeonGreen fluorescent protein....

10.1038/srep34447 article EN cc-by Scientific Reports 2016-09-28

This document presents a detailed description of the challenge on clarifying questions for dialogue systems (ClariQ). The is organized as part Conversational AI series (ConvAI3) at Search Oriented (SCAI) EMNLP workshop in 2020. main aim conversational to return an appropriate answer response user requests. However, some requests might be ambiguous. In IR settings such situation handled mainly thought diversification search result page. It however much more challenging with limited bandwidth....

10.48550/arxiv.2009.11352 preprint EN other-oa arXiv (Cornell University) 2020-01-01

Transformer-based models show their effectiveness across multiple domains and tasks. The self-attention allows to combine information from all sequence elements into context-aware representations. However, global local has be stored mostly in the same element-wise Moreover, length of an input is limited by quadratic computational complexity self-attention. In this work, we propose study a memory-augmented segment-level recurrent Transformer (RMT). Memory store process as well pass between...

10.48550/arxiv.2207.06881 preprint EN other-oa arXiv (Cornell University) 2022-01-01

Multi-agent reinforcement learning (MARL) demonstrates significant progress in solving cooperative and competitive multi-agent problems various environments. One of the principal challenges MARL is need for explicit prediction agents' behavior to achieve cooperation. To resolve this issue, we propose Shared Recurrent Memory Transformer (SRMT) which extends memory transformers settings by pooling globally broadcasting individual working memories, enabling agents exchange information...

10.48550/arxiv.2501.13200 preprint EN arXiv (Cornell University) 2025-01-22

A range of recent works addresses the problem compression sequence tokens into a shorter real-valued vectors to be used as inputs instead token embeddings or key-value cache. These approaches allow reduce amount compute in existing language models. Despite relying on powerful models encoders, maximum attainable lossless ratio is typically not higher than x10. This fact highly intriguing because, theory, information capacity large far beyond presented rates even for 16-bit precision and...

10.48550/arxiv.2502.13063 preprint EN arXiv (Cornell University) 2025-02-18

ABSTRACT The advent of advanced sequencing technologies has significantly reduced the cost and increased feasibility assembling high-quality genomes. Yet, annotation genomic elements remains a complex challenge. Even for species with comprehensively annotated reference genomes, functional assessment individual genetic variants is not straightforward. In response to these challenges, recent breakthroughs in machine learning have led development DNA language models. These transformer-based...

10.1101/2024.04.26.591391 preprint EN bioRxiv (Cold Spring Harbor Laboratory) 2024-04-29

Artem Vazhentsev, Gleb Kuzmin, Shelmanov, Akim Tsvigun, Evgenii Tsymbalov, Kirill Fedyanin, Maxim Panov, Alexander Panchenko, Gusev, Mikhail Burtsev, Manvel Avetisian, Leonid Zhukov. Proceedings of the 60th Annual Meeting Association for Computational Linguistics (Volume 1: Long Papers). 2022.

10.18653/v1/2022.acl-long.566 article EN cc-by Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) 2022-01-01

This paper addresses the challenge of processing long documents using generative transformer models. To evaluate different approaches, we introduce BABILong, a new benchmark designed to assess model capabilities in extracting and distributed facts within extensive texts. Our evaluation, which includes benchmarks for GPT-4 RAG, reveals that common methods are effective only sequences up $10^4$ elements. In contrast, fine-tuning GPT-2 with recurrent memory augmentations enables it handle tasks...

10.48550/arxiv.2402.10790 preprint EN arXiv (Cornell University) 2024-02-16

We propose a general scheme of intelligent adaptive control system based on the Petr K. Anokhin's theory functional systems. This is aimed at controlling purposeful behavior an animat (a simulated animal) that has several natural needs (e.g., energy replenishment, reproduction). The consists set hierarchically linked systems and enables predictive goal-directed behavior. Each includes neural network critic design. also discuss schemes prognosis, decision making, action selection learning...

10.1109/ijcnn.2004.1380879 article EN 2005-01-31

The aim of [email protected] is to bring together IR and AI communities instigate future direction search-oriented conversational systems. We identified the number research areas related which actual interest both have not been fully explored yet. think it's beneficial exchange our visions. solicit paper submissions more importantly proposals for panel discussions where researchers can opinions experiences. believe that proposed workshop relevant ICTIR since we look novel contributions...

10.1145/3121050.3121111 article EN 2017-09-29

Diliara Zharikova, Daniel Kornev, Fedor Ignatov, Maxim Talimanchuk, Dmitry Evseev, Ksenya Petukhova, Veronika Smilga, Karpov, Yana Shishkina, Kosenko, Mikhail Burtsev. Proceedings of the 61st Annual Meeting Association for Computational Linguistics (Volume 3: System Demonstrations). 2023.

10.18653/v1/2023.acl-demo.58 article EN cc-by 2023-01-01
Coming Soon ...