The datasets, ML models, and documentation for monitoring are all available in open access. Such a set will help test hypotheses that detect, segment, and classify objects in various scientific projects. In particular, developers will be able to identify problems in the balance of microorganisms in bodies of water and, as a result, help protect them as best as possible.
Yandex Cloud releases neural network for monitoring Lake Baikal ecosystem in open source
Along with a documentary about the project.
December 20, 2022
5 mins to read
Yandex Cloud published the source code of its ML algorithm used to monitor conditions in Lake Baikal. This technology helps analyze water samples, to then identify and classify the microorganisms present in each one. Biologists and ML specialists around the world can now use the neural network to develop their own systems to monitor bodies of water and track important environmental trends. The algorithm is published on GitHub under the Apache 2.0 Open License.
Initially, the neural network was developed for the “Point No. 1” project, which has been active since 1945. Scientists of the Irkutsk State University’s Research Institute of Biology have been manually analyzing the phyto- and zooplankton of Lake Baikal for more than 75 years to monitor the state of the lake. The neural network has helped automate and accelerate the entire research cycle. Now the algorithm is able to work with 70 forms of plankton. It continues to train other algorithms in the cloud service for the further development of Yandex DataSphere’s ML algorithms. The MaritimeAI company and the Lake Baikal Foundation for the Support of Applied Environmental Research and Development also participated in the creation of this technology.
The documentary “12,000 Samples” tells about how scientists and developers have gone from solving a local problem in Lake Baikal to developing a neural network in open source. The plot is based on conversations with specialists of the Biological Research Institute, MaritimeAI developers, and Yandex employees. The participants told why it is sometimes difficult for scientists and IT professionals to find a common language, how they struggled with the project’s difficulties, and how cloud technologies and AI can solve scientific problems. The film can be viewed for free on the Yandex YouTube channel and on the KinoPoisk service.
Learn more about Yandex DataSphere
“Technological accessibility is one of our key priorities. Yandex not only regularly releases its solutions in open source, but also helps develop open-source code in joint projects with third-party developers. Yandex is the leader among Russian companies in terms of contributions to open-source projects, according to OSS Index. By publishing the ecological monitoring algorithm with MaritimeAI is another example of Yandex as a company sticking to its own principles. The development will help not just to solve the problem of a specific group of scientists, but to test hypotheses about the water conditions anywhere and everywhere. After all, Baikal is far from the only place on the planet where such monitoring is conducted,” said Alexey Bashkeev, head of the Yandex Cloud platform.
