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

  Jin S., Komarov M. M.

  Reproducible Benchmark of Wavelet-Enhanced Intrabody Communication Biometric Identification

  Intrabody communication (IBC) channels offer physiological diversity that can be leveraged for passive biometric identification in wearable devices. Recent reports of over 99 per cent identification accuracy have frequently resulted from data leakage, where samples from the same subject are seen in both training and evaluation, yielding inflated and unreliable metrics. In this work, we establish a public, leakage-free benchmark for IBC biometrics built on a 30-subject open dataset, using strict subject-wise 80/20 splits repeated five times to ensure reproducibility. We systematically compare frequency-domain and time-frequency representations, including resampled spectra, discrete wavelet transform (DWT) statistics, and their fusion. We evaluate a diverse suite of models, spanning k-nearest neighbors, Random Forest, a one-dimensional spectral convolutional neural network (SpectralCNN), a multilayer perceptron (MLP) for feature fusion, and an SVD-based linear decoder. Our strongest configuration – an MLP trained on combined resampled spectra and DWT features – attains an accuracy of approximately 83 per cent, which serves as an upper bound under our protocol, substantially outperforming classical instancebased methods (approx. 41 per cent) and SVD (approx. 42 per cent). The SpectralCNN, trained on resampled spectra alone, achieves 74 per cent accuracy. Confusion matrix analysis reveals that residual errors are concentrated among subject pairs with statistically overlapping signatures, suggesting the presence of intrinsically hard users and a potential biometric ceiling for this modality. Embedded profiling on an STM32F446RE Cortex-M4 microcontroller indicates that lifting-based wavelet features enable low-latency, low-energy scoring, requiring approximately 0.55 ms and 18 micro-J per 256-point spectrum for Lift-bior feature extraction plus Random Forest inference (versus approx. 33 micro-J for the equivalent db4-DWT pipeline). All code, data split scripts, and Jupyter notebooks are released open source to facilitate reproducibility and enable rigorous future comparisons.

  Scientific Reports. 2026.

• Book chapter

  Alexander Baranov, Anna Palatkina, Makovka Y. et al.

  KoWit-24: A Richly Annotated Dataset of Wordplay in News Headlines

  We present KoWit-24, a dataset with fine-grained annotation of wordplay in 2,700 Russian news headlines. KoWit-24 annotations include the presence of wordplay, its type, wordplay anchors, and words/phrases the wordplay refers to. Unlike the majority of existing humor collections of canned jokes, KoWit-24 provides wordplay contexts – each headline is accompanied by the news lead and summary. The most common type of wordplay in the dataset is the transformation of collocations, idioms, and named entities – the mechanism that has been underrepresented in previous humor datasets. Our experiments with five LLMs show that there is ample room for improvement in wordplay detection and interpretation tasks. The dataset and evaluation scripts are available at https://github.com/Humor-Research/KoWit-24

  In bk.: Proceedings of the 15th International Conference on Recent Advances in Natural Language Processing. Shumen: INCOMA Ltd, 2025. P. 125-132.

• 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