Dr. Goharik Petrosyan Razmik is a Senior Researcher at the Schmidt Institute of Physics of the Earth of the Russian Academy of Sciences and an Assistant Professor at the International Scientific and Educational Center of the National Academy of Sciences of Armenia. She earned her Ph.D. in Physics and Mathematics with a specialization in Mathematical Modeling, Numerical Methods, and Software Systems, and holds a master's degree in Applied Mathematics with honors from Yerevan State University. With more than two decades of academic, research, and industry experience, Dr. Petrosyan has taught a wide range of subjects including mathematical modeling, numerical methods, probability theory, linear algebra, programming, and electronic business. Her research interests encompass applied mathematics, artificial intelligence, bioinformatics, mathematical modeling in medicine, economic and mathematical modeling, seismology, formal languages, and distributed systems. She is an active member of the Armenian Mathematical Union and has contributed significantly to interdisciplinary research and higher education in Armenia and Russia.

Analysis of Seismic Activity in the Republic of Armenia and Adjacent Territories (2008-2024) Based on Public Catalogs Using AI In seismology, the quality of the final conclusions is determined by the characteristics of the seismic catalogs used—primarily the completeness, accuracy, and reliability of the underlying data. Earthquakes for the territory of Armenia are mainly described on the basis of instrumental observations. Preliminary assessments reported by different researchers indicate that, at the present stage, the quality of Armenian bulletins and the associated catalogs is insufficient to fully extract and utilize the information contained in them for contemporary geophysical applications. A substantial contribution to this field can be provided by artificial intelligence: AI can detect and account for latent data heterogeneity by analyzing not only the events themselves, but also the conditions under which they are recorded. As noted in several publications, the original seismological information presented in bulletins and catalogs is characterized by pronounced spatial non-uniformity, generally attributable to differences in the density of the seismic station network and to uneven geographic coverage of the investigated region.  ObjectivesThis study is aimed at a comprehensive analysis of seismicity across the Republic of Armenia and adjacent areas for the period 2008–2024, using publicly available catalogs and applying artificial intelligence methods. The main objectives are as follows: (i) identification of spatiotemporal trends; (ii) assessment of the completeness and consistency of open catalogs; (iii) demonstration of AI’s capabilities for automated event classification, detection of data anomalies, and forecasting parameters of seismic activity over short- and medium-term time horizons. The use of illustrative materials and visualizations is intended to improve the clarity of the presented results and to strengthen the supporting argumentation. Approach Open earthquake catalogs and related station data from the Western Caucasus and adjacent regions are used. The datasets are processed and standardized. Modern machine learning methods are applied: event classification based on physical parameters (magnitude, depth, source mechanism), anomaly detection, and regression models to approximate recurrence behavior and the probability of seismic events. The visualization suite includes dynamic maps, time-evolution plots, and pattern-discrimination diagrams generated with AI tools to improve interpretability.  Findings The study is expected to reveal robust spatiotemporal regularities. The AI models should demonstrate high discriminative performance in classifying events and in detecting early anomalies without being dependent on arbitrary discretization of catalog data. A key outcome will be the identification of areas with poorly constrained parameters and recommendations for improving catalog reliability and consistency. The principal limitations are associated with incomplete spatial-temporal coverage and heterogeneity in detection and recording practices across catalogs.  Conclusion The integration of open seismic catalogs with AI-driven analytical methods provides a more accurate characterization of regional seismicity and supports adaptive monitoring and risk assessment strategies for Armenia and surrounding territories. The work highlights the potential of automated data processing and visualization for both scientific and applied purposes.  Keywords: seismic activity; Armenia; open-access catalogs; artificial intelligence; spatiotemporal dynamics; earthquake risk.