Computer modeling of natural Systems (NSS Lab)
Publications
55
Citations
416
h-index
10
Authorization required.
Research in the laboratory focuses on several main areas. Algorithmic directions: automatic machine learning, generative design of physical objects, algorithms for identifying physical models based on data, algorithms for learning probabilistic models, as well as algorithms for composite AI that can combine all of the above :)
Modeling of natural processes has been and remains a favorite applied field of our laboratory. In addition, the laboratory pays a lot of attention to the development of open source and the creation of modern educational courses in the master's degree.
- Machine learning
- Generative design of physical objects
- AutoML
- Algorithms for identifying physical models based on data
- Composite AI
- Bayesian networks
Anna Kalyuzhnaya
Head of Laboratory
Ilia Revin
Researcher
Irina Deeva
Researcher
Anna Bubnova
Engineer
Yulia Borisova
Engineer
Mikhail Maslyaev
Junior researcher
Mikhail Sarafanov
Junior researcher
Research directions
Other projects
+
We still have many interesting small student or initiative projects. You can get acquainted with the zoo of our developments here (https://github.com/ITMO-NSS-team )
Generative design of physical objects
+
We are developing a library for generative design of geometric objects coupled with a continuous GEFEST environment (https://github.com/ITMO-NSS-team/GEFEST ).
AI for hydrometeorological tasks
+
We are trying to combine our knowledge in the field of hydromet modeling with the capabilities of machine learning and intelligent optimization methods to improve the quality of modeling natural processes.
AI for oil industry tasks
+
Recently, we have been solving many industrial problems using AI methods, primarily in the oil industry.
Identification of the DP by data
+
We are developing a framework for restoring the structure of the DP according to EPDE data (https://github.com/ITMO-NSS-team/EPDE )
Learning algorithms for Bayesian networks
+
We are developing an open library for structural and parametric training of Bayesian BAMT networks (https://github.com/ITMO-NSS-team/BAMT )
Automatic machine learning
+
We are developing an open framework for automatic machine learning FEDOT (https://github.com/nccr-itmo/FEDOT )
Publications and patents
Partners
Lab address
Санкт-Петербург, Кронверкский просп., 49
Authorization required.