Centre for Language and Semantic Technologies

Publications

• Book

  Kuznetsov S.

  Data Analytics and Management in Data Intensive Domains: 25th International Conference, DAMDID/RCDL 2023, Moscow, Russia, October 24–27, 2023, Revised Selected Papers

  This book constitutes the post-conference proceedings of the 25th International Conference on Data Analytics and Management in Data Intensive Domains, DAMDID/RCDL 2023, held in Moscow, Russia, during 24-27 October 2023.

  
  The 21 papers presented here were carefully reviewed and selected from 75 submissions. These papers are organized in the following topical sections: Data Models and Knowledge Graphs; Databases in Data Intensive Domains; Machine learning methods and applications; Data Analysis in Astronomy & Information extraction from text. Papers from keynote talks have also been included in this book.

   

  
  

  Vol. 2086: Communications in Computer and Information Science. Springer, 2024.

• Article

  - Р. М., Koltcov Sergei, Surkov A. et al.

  Modeling Pruning as a Phase Transition: A Thermodynamic Analysis of Neural Activations

  Activation pruning reduces neural network complexity by eliminating low-importance neuron activations, yet identifying the critical pruning threshold—beyond which accuracy rapidly deteriorates—remains computationally expensive and typically requires exhaustive search. We introduce a thermodynamics-inspired framework that treats activation distributions as energy-filtered physical systems and employs the free energy of activations as a principled evaluation metric. Phase-transition–like phenomena in the free-energy profile—such as extrema, inflection points, and curvature changes—yield reliable estimates of the critical pruning threshold, providing a theoretically grounded means of predicting sharp accuracy degradation. To further enhance efficiency, we propose a renormalized free energy technique that approximates full-evaluation free energy using only the activation distribution of the unpruned network. This eliminates repeated forward passes, dramatically reducing computational overhead and achieving speedups of up to 550× for MLPs. Extensive experiments across diverse vision architectures (MLP, CNN, ResNet, MobileNet, Vision Transformer) and text models (LSTM, BERT, ELECTRA, T5, GPT-2) on multiple datasets validate the generality, robustness, and computational efficiency of our approach. Overall, this work establishes a theoretically grounded and practically effective framework for activation pruning, bridging the gap between analytical understanding and efficient deployment of sparse neural networks.

  Computers, Materials and Continua. 2025. P. 1-24.

• Book chapter

  Dudyrev E., Couceiro M., Kaytoue M. et al.

  Atomic Patterns for Efficient Computation with Pattern Structures

  Pattern Structures is a framework in FCA allowing objects to have complex descriptions, only requiring that the set of descriptions forms a complete meet-semi-lattice. However, some particular descrip tions or patterns, such as subgraphs and subsequences, do not necessarily ensure that every pair of descriptions has a unique infimum and ask for additional operations, e.g., anti-chain completion. Moreover, meet-based approaches struggle to generate non-trivial implications for complex data since, in general, they only output closed descriptions. For overcoming such limitations, we introduce in this paper an alternative view of pat tern structures based on the join operation and the so-called “atomic patterns”. Such atomic patterns correspond to join-irreducible descrip tions in the join-semi-lattice of all possible descriptions. They enable an efficient traversal of the description space and the computation of closures, minimal generators, pseudo-intents, implications among others, while showing very good computational performance.

  In bk.: Second International Joint Conference, CONCEPTS 2025, Cluj-Napoca, Romania, September 8–12, 2025, Proceedings. Conceptual Knowledge Structures. LNCS, volume 15941. Cham: Springer, 2025. P. 178-194.

• Working paper

  Меньшиков И. А., Бернадотт А. К., Elvimov N. S.

  Hessian-based lightweight neural network for brain vessel segmentation on a minimal training dataset

  Accurate segmentation of blood vessels in brain magnetic resonance angiography (MRA) is essential for successful surgical procedures, such as aneurysm repair or bypass surgery. Currently, annotation is primarily performed through manual segmentation or classical methods, such as the Frangi filter, which often lack sufficient accuracy. Neural networks have emerged as powerful tools for medical image segmentation, but their development depends on well-annotated training datasets. However, there is a notable lack of publicly available MRA datasets with detailed brain vessel annotations. To address this gap, we propose a novel semi-supervised learning lightweight neural network with Hessian matrices on board for 3D segmentation of complex structures such as tubular structures, which we named HessNet. The solution is a Hessian-based neural network with only 6000 parameters. HessNet can run on the CPU and significantly reduces the resource requirements for training neural networks. The accuracy of vessel segmentation on a minimal training dataset reaches state-of-the-art results. It helps us create a large, semi-manually annotated brain vessel dataset of brain MRA images based on the IXI dataset (annotated 200 images). Annotation was performed by three experts under the supervision of three neurovascular surgeons after applying HessNet. It provides high accuracy of vessel segmentation and allows experts to focus only on the most complex important cases. The dataset is available at https://git.scinalytics.com/terilat/VesselDatasetPartly.

  Statistical mechanics. arXie. arXive, 2025