arXiv subjects

Oleg Smirnov

Publications and source records attributed to Oleg Smirnov.

At least 37 records · Page 2Linked to original sources

arXiv2025

Are We Really Measuring Progress? Transferring Insights from Evaluating Recommender Systems to Temporal Link Prediction

Recent work has questioned the reliability of graph learning benchmarks, citing concerns around task design, methodological rigor, and data suitability. In this extended abstract, we contribute to this discussion by focusing on evaluation strategies in Temporal Link Prediction (TLP). We observe that current evaluation protocols are often affected by one or more of the following issues: (1) inconsistent sampled metrics, (2) reliance on hard negative sampling often introduced as a means to improve robustness, and (3) metrics that implicitly assume equal base probabilities across source nodes by combining predictions. We support these claims through illustrative examples and connections to longstanding concerns in the recommender systems community. Our ongoing work aims to systematically characterize these problems and explore alternatives that can lead to more robust and interpretable evaluation. We conclude with a discussion of potential directions for improving the reliability of TLP benchmarks.

cs.LG↗

arXiv2025

Expressivity of Representation Learning on Continuous-Time Dynamic Graphs: An Information-Flow Centric Review

Graphs are ubiquitous in real-world applications, ranging from social networks to biological systems, and have inspired the development of Graph Neural Networks (GNNs) for learning expressive representations. While most research has centered on static graphs, many real-world scenarios involve dynamic, temporally evolving graphs, motivating the need for Continuous-Time Dynamic Graph (CTDG) models. This paper provides a comprehensive review of Graph Representation Learning (GRL) on CTDGs with a focus on Self-Supervised Representation Learning (SSRL). We introduce a novel theoretical framework that analyzes the expressivity of CTDG models through an Information-Flow (IF) lens, quantifying their ability to propagate and encode temporal and structural information. Leveraging this framework, we categorize existing CTDG methods based on their suitability for different graph types and application scenarios. Within the same scope, we examine the design of SSRL methods tailored to CTDGs, such as predictive and contrastive approaches, highlighting their potential to mitigate the reliance on labeled data. Empirical evaluations on synthetic and real-world datasets validate our theoretical insights, demonstrating the strengths and limitations of various methods across long-range, bi-partite and community-based graphs. This work offers both a theoretical foundation and practical guidance for selecting and developing CTDG models, advancing the understanding of GRL in dynamic settings.

cs.LG↗

arXiv2025

On the Power of Heuristics in Temporal Graphs

Dynamic graph datasets often exhibit strong temporal patterns, such as recency, which prioritizes recent interactions, and popularity, which favors frequently occurring nodes. We demonstrate that simple heuristics leveraging only these patterns can perform on par or outperform state-of-the-art neural network models under standard evaluation protocols. To further explore these dynamics, we introduce metrics that quantify the impact of recency and popularity across datasets. Our experiments on BenchTemp and the Temporal Graph Benchmark show that our approaches achieve state-of-the-art performance across all datasets in the latter and secure top ranks on multiple datasets in the former. These results emphasize the importance of refined evaluation schemes to enable fair comparisons and promote the development of more robust temporal graph models. Additionally, they reveal that current deep learning methods often struggle to capture the key patterns underlying predictions in real-world temporal graphs. For reproducibility, we have made our code publicly available.

cs.LG↗

arXiv2024

Understanding Players as if They Are Talking to the Game in a Customized Language: A Pilot Study

This pilot study explores the application of language models (LMs) to model game event sequences, treating them as a customized natural language. We investigate a popular mobile game, transforming raw event data into textual sequences and pretraining a Longformer model on this data. Our approach captures the rich and nuanced interactions within game sessions, effectively identifying meaningful player segments. The results demonstrate the potential of self-supervised LMs in enhancing game design and personalization without relying on ground-truth labels.

cs.LG↗

arXiv2024

player2vec: A Language Modeling Approach to Understand Player Behavior in Games

Methods for learning latent user representations from historical behavior logs have gained traction for recommendation tasks in e-commerce, content streaming, and other settings. However, this area still remains relatively underexplored in video and mobile gaming contexts. In this work, we present a novel method for overcoming this limitation by extending a long-range Transformer model from the natural language processing domain to player behavior data. We discuss specifics of behavior tracking in games and propose preprocessing and tokenization approaches by viewing in-game events in an analogous way to words in sentences, thus enabling learning player representations in a self-supervised manner in the absence of ground-truth annotations. We experimentally demonstrate the efficacy of the proposed approach in fitting the distribution of behavior events by evaluating intrinsic language modeling metrics. Furthermore, we qualitatively analyze the emerging structure of the learned embedding space and show its value for generating insights into behavior patterns to inform downstream applications.

cs.LG↗

arXiv2024

Behavior Structformer: Learning Players Representations with Structured Tokenization

In this paper, we introduce the Behavior Structformer, a method for modeling user behavior using structured tokenization within a Transformer-based architecture. By converting tracking events into dense tokens, this approach enhances model training efficiency and effectiveness. We demonstrate its superior performance through ablation studies and benchmarking against traditional tabular and semi-structured baselines. The results indicate that structured tokenization with sequential processing significantly improves behavior modeling.

cs.CL↗

arXiv2024

Kotlin ML Pack: Technical Report

In this technical report, we present three novel datasets of Kotlin code: KStack, KStack-clean, and KExercises. We also describe the results of fine-tuning CodeLlama and DeepSeek models on this data. Additionally, we present a version of the HumanEval benchmark rewritten by human experts into Kotlin - both the solutions and the tests. Our results demonstrate that small, high-quality datasets (KStack-clean and KExercises) can significantly improve model performance on code generation tasks, achieving up to a 16-point increase in pass rate on the HumanEval benchmark. Lastly, we discuss potential future work in the field of improving language modeling for Kotlin, including the use of static analysis tools in the learning process and the introduction of more intricate and realistic benchmarks.

cs.SE↗

arXiv2023

A Wideband Polarization Observation of Hydra A with the Jansky Very Large Array

We present results of a wideband high-resolution polarization study of Hydra A, one of the most luminous FR I radio galaxies known and amongst the most well-studied. The radio emission from this source displays extremely large Faraday rotation measures (RM), ranging from -12300 rad m$^{-2}$ to 5000 rad m$^{-2}$, the majority of which are believed to originate from magnetized thermal gas external to the radio tails. The radio emission from both tails strongly depolarizes with decreasing frequency. The depolarization, as a function of wavelength, is commonly non-monotonic, often showing oscillatory behavior, with strongly non-linear rotation of the polarization position angle with $λ^2$. A simple model, based on the RM screen derived from the high frequency, high resolution data, predicts the lower frequency depolarization remarkably well. The success of this model indicates the majority of the depolarization can be attributed to fluctuations in the magnetic field on scales $< 1500$ pc, suggesting the presence of turbulent magnetic field/electron density structures on sub-kpc scales within a Faraday rotating (FR) medium.

astro-ph.GA↗

arXiv2023

On the use of temporal filtering for mitigating galactic synchrotron calibration bias in 21 cm reionization observations

Precision antenna calibration is required for mitigating the impact of foreground contamination in 21 cm cosmological radio surveys. One widely studied source of error is the effect of missing point sources in the calibration sky model; however, poorly understood diffuse galactic emission also creates a calibration bias that can complicate the clean separation of foregrounds from the 21 cm signal. In this work, we present a technique for suppressing this bias with temporal filtering of radio interferometric visibilities observed in a drift-scan mode. We demonstrate this technique on mock simulations of the Hydrogen Epoch of Reionization Array (HERA) experiment. Inspecting the recovered calibration solutions, we find that our technique reduces spurious errors by over an order of magnitude. This improved accuracy approaches the required accuracy needed to make a fiducial detection of the 21 cm signal with HERA, but is dependent on a number of external factors that we discuss. We also explore different types of temporal filtering techniques and discuss their relative performance and tradeoffs.

astro-ph.CO↗

arXiv2022

Rooms with Text: A Dataset for Overlaying Text Detection

In this paper, we introduce a new dataset of room interior pictures with overlaying and scene text, totalling to 4836 annotated images in 25 product categories. We provide details on the collection and annotation process of our dataset, and analyze its statistics. Furthermore, we propose a baseline method for overlaying text detection, that leverages the character region-aware text detection framework to guide the classification model. We validate our approach and show its efficiency in terms of binary classification metrics, reaching the final performance of 0.95 F1 score, with false positive and false negative rates of 0.02 and 0.06 correspondingly.

cs.CV↗

arXiv2022

Potential for a precision measurement of solar $pp$ neutrinos in the Serappis Experiment

The Serappis (SEarch for RAre PP-neutrinos In Scintillator) project aims at a precision measurement of the flux of solar $pp$ neutrinos on the few-percent level. Such a measurement will be a relevant contribution to the study of solar neutrino oscillation parameters and a sensitive test of the solar luminosity constraint. The concept of Serappis relies on a small organic liquid scintillator detector ($\sim$20 m$^3$) with excellent energy resolution ($\sim$2.5 % at 1 MeV), low internal background and sufficient shielding from surrounding radioactivity. This can be achieved by a minor upgrade of the OSIRIS facility at the site of the JUNO neutrino experiment in southern China. To go substantially beyond current accuracy levels for the $pp$ flux, an organic scintillator with ultra-low $^{14}$C levels (below $10^{-18}$) is required. The existing OSIRIS detector and JUNO infrastructure will be instrumental in identifying suitable scintillator materials, offering a unique chance for a low-budget high-precision measurement of a fundamental property of our Sun that will be otherwise hard to access.

physics.ins-det↗

arXiv2022

WS-Snapshot: An effective algorithm for wide-field and large-scale imaging

The Square Kilometre Array (SKA) is the largest radio interferometer under construction in the world. The high accuracy, wide-field and large size imaging significantly challenge the construction of the Science Data Processor (SDP) of SKA. We propose a hybrid imaging method based on improved W-Stacking and snapshots. The w range is reduced by fitting the snapshot $uv$ plane, thus effectively enhancing the performance of the improved W-Stacking algorithm. We present a detailed implementation of WS-Snapshot. With full-scale SKA1-LOW simulations, we present the imaging performance and imaging quality results for different parameter cases. The results show that the WS-Snapshot method enables more efficient distributed processing and significantly reduces the computational time overhead within an acceptable accuracy range, which would be crucial for subsequent SKA science studies.

astro-ph.IM↗

arXiv2022

IntelliTC: Automating Type Changes in IntelliJ IDEA

Developers often change types of program elements. Such refactoring often involves updating not only the type of the element itself, but also the API of all type-dependent references in the code, thus it is tedious and time-consuming. Despite type changes being more frequent than renamings, just a few current IDE tools provide partially-automated support only for a small set of hard-coded types. Researchers have recently proposed a data-driven approach to inferring API rewrite rules for type change patterns in Java using code commits history. In this paper, we build upon these recent advances and introduce IntelliTC - a tool to perform Java type change refactoring. We implemented it as a plugin for IntelliJ IDEA, a popular Java IDE developed by JetBrains. We present 3 different ways of providing support for such a refactoring from the standpoint of the user experience: Classic mode, Suggested Refactoring, and Inspection mode. To evaluate these modalities of using IntelliTC, we surveyed 22 experienced software developers. They positively rated the usefulness of the tool. The source code and distribution of the plugin are available on GitHub: https://github.com/JetBrains-Research/data-driven-type-migration. A demonstration video is on YouTube: https://youtu.be/pdcfvADA1PY.

cs.SE↗

arXiv2022

LADUMA: Discovery of a luminous OH megamaser at $z > 0.5$

In the local Universe, OH megamasers (OHMs) are detected almost exclusively in infrared-luminous galaxies, with a prevalence that increases with IR luminosity, suggesting that they trace gas-rich galaxy mergers. Given the proximity of the rest frequencies of OH and the hyperfine transition of neutral atomic hydrogen (HI), radio surveys to probe the cosmic evolution of HI in galaxies also offer exciting prospects for exploiting OHMs to probe the cosmic history of gas-rich mergers. Using observations for the Looking At the Distant Universe with the MeerKAT Array (LADUMA) deep HI survey, we report the first untargeted detection of an OHM at $z > 0.5$, LADUMA J033046.20$-$275518.1 (nicknamed "Nkalakatha"). The host system, WISEA J033046.26$-$275518.3, is an infrared-luminous radio galaxy whose optical redshift $z \approx 0.52$ confirms the MeerKAT emission line detection as OH at a redshift $z_{\rm OH} = 0.5225 \pm 0.0001$ rather than HI at lower redshift. The detected spectral line has 18.4$σ$ peak significance, a width of $459 \pm 59\,{\rm km\,s^{-1}}$, and an integrated luminosity of $(6.31 \pm 0.18\,{\rm [statistical]}\,\pm 0.31\,{\rm [systematic]}) \times 10^3\,L_\odot$, placing it among the most luminous OHMs known. The galaxy's far-infrared luminosity $L_{\rm FIR} = (1.576 \pm 0.013) \times 10^{12}\,L_\odot$ marks it as an ultra-luminous infrared galaxy; its ratio of OH and infrared luminosities is similar to those for lower-redshift OHMs. A comparison between optical and OH redshifts offers a slight indication of an OH outflow. This detection represents the first step towards a systematic exploitation of OHMs as a tracer of galaxy growth at high redshifts.

astro-ph.GA↗

arXiv2022

RFI Identification Based On Deep-Learning]{A Robust RFI Identification For Radio Interferometry based on a Convolutional Neural Network

The rapid development of new generation radio interferometers such as the Square Kilometer Array (SKA) has opened up unprecedented opportunities for astronomical research. However, anthropogenic Radio Frequency Interference (RFI) from communication technologies and other human activities severely affects the fidelity of observational data. It also significantly reduces the sensitivity of the telescopes. We proposed a robust Convolutional Neural Network (CNN) model to identify RFI based on machine learning methods. We overlaid RFI on the simulation data of SKA1-LOW to construct three visibility function datasets. One dataset was used for modeling, and the other two were used for validating the model's usability. The experimental results show that the Area Under the Curve (AUC) reaches 0.93, with satisfactory accuracy and precision. We then further investigated the effectiveness of the model by identifying the RFI in the actual observational data from LOFAR and MeerKAT. The results show that the model performs well. The overall effectiveness is comparable to AOFlagger software and provides an improvement over existing methods in some instances.

astro-ph.IM↗

arXiv2022

MeqSilhouette v2: Spectrally-resolved polarimetric synthetic data generation for the Event Horizon Telescope

We present MeqSilhouette v2.0 (MeqSv2), a fully polarimetric, time-and frequency-resolved synthetic data generation software for simulating millimetre (mm) wavelength very long baseline interferometry (VLBI) observations with heterogeneous arrays. Synthetic data are a critical component in understanding real observations, testing calibration and imaging algorithms, and predicting performance metrics of existing or proposed sites. MeqSv2 applies physics-based instrumental and atmospheric signal corruptions constrained by empirically-derived site and station parameters to the data. The new version is capable of applying instrumental polarization effects and various other spectrally-resolved effects using the Radio Interferometry Measurement Equation (RIME) formalism and produces synthetic data compatible with calibration pipelines designed to process real data. We demonstrate the various corruption capabilities of MeqSv2 using different arrays, with a focus on the effect of complex bandpass gains on closure quantities for the EHT at 230 GHz. We validate the frequency-dependent polarization leakage implementation by performing polarization self-calibration of synthetic EHT data using PolSolve. We also note the potential applications for cm-wavelength VLBI array analysis and design and future directions.

astro-ph.IM↗

arXiv2022

Performance evaluation of baseline-dependent averaging based onfull-scale SKA1-LOW simulation

The Square Kilometre Array (SKA) is the largest radio interferometer under construction in the world. Its immense amount of visibility data poses a considerable challenge to the subsequent processing by the science data processor (SDP). Baseline dependent averaging (BDA), which reduces the amount of visibility data based on the baseline distribution of the radio interferometer, has become a focus of SKA SDP development. This paper developed and implemented a full-featured BDA module based on Radio Astronomy Simulation, Calibration and Imaging Library (RASCIL). Simulated observations were then performed with RASCIL based on a full-scale SKA1-LOW configuration. The performance of the BDA was systematically investigated and evaluated based on the simulated data. The experimental results presented that the amount of visibility data is reduced by about 50\% to 85\% for different time intervals ($Δt_{max}$). In addition, different $Δt_{max}$ have a significant effect on the imaging quality. The smaller the $Δt_{max}$, the smaller the degradation of the imaging quality.

astro-ph.IM↗

arXiv2021

The Panchromatic Afterglow of GW170817: The full uniform dataset, modeling, comparison with previous results and implications

We present the full panchromatic afterglow light curve data of GW170817, including new radio data as well as archival optical and X-ray data, between 0.5 and 940 days post-merger. By compiling all archival data, and reprocessing a subset of it, we have evaluated the impact of differences in data processing or flux determination methods used by different groups, and attempted to mitigate these differences to provide a more uniform dataset. Simple power-law fits to the uniform afterglow light curve indicate a $t^{0.86\pm0.04}$ rise, a $t^{-1.92\pm0.12}$ decline, and a peak occurring at $155\pm4$ days. The afterglow is optically thin throughout its evolution, consistent with a single spectral index ($-0.584\pm0.002$) across all epochs. This gives a precise and updated estimate of the electron power-law index, $p=2.168\pm0.004$. By studying the diffuse X-ray emission from the host galaxy, we place a conservative upper limit on the hot ionized ISM density, $<$0.01 cm$^{-3}$, consistent with previous afterglow studies. Using the late-time afterglow data we rule out any long-lived neutron star remnant having magnetic field strength between 10$^{10.4}$ G and 10$^{16}$ G. Our fits to the afterglow data using an analytical model that includes VLBI proper motion from Mooley et al. (2018), and a structured jet model that ignores the proper motion, indicates that the proper motion measurement needs to be considered while seeking an accurate estimate of the viewing angle.

astro-ph.HE↗