KOALA: A Kalman Optimization Algorithm with Loss Adaptivity
A. Davtyan, S. Sameni, L. Cerkezi, G. Meishvili, A. Bielski and P. Favaro. "KOALA: A Kalman Optimization Algorithm with Loss Adaptivity." In AAAI 2022.
Curriculum Vitae
Education
PhD in Computer Science - University of Bern, 2020-present
Topics: Learning object interactions, Video generation
Advisor: Prof. Paolo FavaroSpecialist Degree in Fundamental Mathematics and Mechanics - Moscow State University, 2020
With specialization in Theory of Functions and Functional Analisys
Thesis Title: Spectral analysis of the operator in the generalized ’t Hooft modelDiploma in Computer Science - Yandex School of Data Science, 2018
With specialization in Data Analisys in Applied Sciences
Thesis Title: Optimization of oil production via construction of forecast models on injection wells operation modes
Professional Experience
Specialist - Inline Group, 2017-2021
Applying machine learning to various problems in oil industry.Software Engineer - Yandex, 2019-2020
Developing the backend of a high-load service (Yandex.Navigator) responsible for building optimal routes in the automobile graph with respect to the realtime traffic. This included implementing algorithms on road graphs and improving quality of the service via supervised and reinforcement learning techniques (C++ and Python).
Publications
Oil production forecast models based on sliding window regression
A. Davtyan, A. Rodin, I. Muchnik and A. Romashkin. "Oil production forecast models based on sliding window regression." In Journal of Petroleum Science and Engineering, Volume 195, 2020, 107916, ISSN 0920-4105, https://doi.org/10.1016/j.petrol.2020.107916.
Skills
Programming:
C/C++, Python, SQL, MapReduce, SVN, git, LaTeXFrameworks:
PyTorch, Tensorflow, SciPy, Numpy, OpenCVLanguages:
English (C1), German (B2), Russian (native), Armenian (native)