Translation. Region: Russian Federal
Source: Novosibirsk State University –
An important disclaimer is at the bottom of this article.
The 53rd IUPAC General Assembly, together with the 50th World Chemical Congress, was held in Kuala Lumpur, Malaysia, on July 12-19. A special feature of the congress was the allocation of a specialized section on “Artificial Intelligence in Chemistry.” Within its framework, Associate Professor of the Department of Solid State Chemistry Faculty of Natural Sciences of NSU, Senior Researcher at the Boreskov Institute of Catalysis SB RAS Anna Nartova gave an invited talk on “Artificial Neural Networks for Microscopic Data Analysis,” in which she presented the iOk online platform to the global scientific community (HTTP: //yok. HSSU.ru/), developed by scientists from Novosibirsk State University as part of the program "Priority 2030".
This platform includes a set of digital services for automatic image analysis using deep machine learning and artificial intelligence methods. Three services are freely available on the platform. Universal Telegram service No Code ML (HTTPS: //t. T.NSU_ML_VOT) is designed for classical training of neural networks on user datasets. The second service is DLgram (HTTPS: //t. TA/NANOPARITICLECLY_NSK) — is designed to recognize multiple homogeneous objects of various natures. Includes training of the neural network by the user using a marked area from the same image. Online service ParticlesNN (HTTP: // Partiklesnn.NSU.ru/) is designed for automatic recognition of nanoparticles in scanning probe microscopy (SPM) and electron microscopy (EM) images by a trained neural network. Cloud services No Code ML and DLgram provide training of a neural network on user objects, automatic recognition of objects in images and the ability to adjust the recognition results by the user. They also analyze the detected objects and determine their parameters, such as quantity, size, area and concentration. The services are able to work with various images – pictures from electron microscopes, photographs from digital cameras (including smartphones). They recognize various objects: nanoparticles, microorganisms, cells, plant seeds, as well as larger objects – animals, plants, various parts, vehicles and much more. At the same time, to work with the services, the user does not need to have any special programming skills or understand neural networks. Preliminary image processing is also not needed. The results are provided in the form of information about all detected objects, and the user can adjust them if necessary. Currently, hundreds of researchers from all over the world use the developed services. At the same time, the services are most in demand by specialists working with images obtained using various types of microscopy.
– In general, the section “Artificial intelligence in chemistry”, despite the fact that the direction only – only the beginning to form and the number of reports was inferior to other, more traditional sections, was perhaps the most popular from the audience. Among the speakers there were many representatives of Russia. Nash the lecture aroused a very great interest in the audience. A lot of listeners went to her purposefully. Such an increased interest in our online platform is due to several reasons, including the fact that we promoted it at several scientific conferences. A number of articles were published, including in high -rated scientific journals. And, I must say that our team, in which A.V. Matveev and A.G. Okunev, was one of the first in the world to present their developments in this direction. Then, single work was devoted to this topic, and we walked alone with them and at times even ahead of them. Now their avalanche is huge, but we still recorded our championship. And even more so, we were the first to offer a “platform approach” when users are offered a ready -made service, and not just a code. At the moment, we have formed our own circle of users, so we received an invitation to make a lecture at this Congress. It was very pleasant to know that our development is known not only in Russia, but also in the world. So, young scientists from China, Malaysia and other countries who were already familiar with our work came to our lecture. During the discussions, after the report, young chemists asked a lot of questions, in most cases, interested in not how neural networks are arranged, but how to use them as a finished tool, while avoiding teaching the Python programming language. They were interested in neural networks, when using which the programmer’s skill is not required. And our platform is just such a product, ”said Anna Nartova.
The section participants separately discussed the problem of the activation barrier, which is of fundamental importance in the use of neural networks in chemistry. It is the need for confident programming skills and deep knowledge of neural networks. According to Anna Nartova, the platform approach is the optimal solution to eliminate it and make neural networks a tool for working in the laboratory. The section participants, working in different areas, also related to chemistry, noted: if a chemist needs to have a high enough level of Python to use a neural network as a tool, he will most likely not do this. The situation should be changed by using a platform approach, which is a ready-made product that does not require the scientist to delve into the code, but allows you to simply upload your data to the application, process it and get a ready result. In this case, the return on using neural networks is maximum.
— We started promoting the idea of implementing a platform approach to solving the problems of accessible use of neural networks for scientists who do not have programming skills at our university several years ago. Now it has become a widespread trend. The first works in the direction of using neural networks in processing microscopic images were aimed at solving some specific cases, and usually the machine learning model was posted in the public domain. In our case, firstly, we are talking about the universalization of the approach, and secondly, about the creation of applications or services that facilitate work with neural networks. Thus, at the same time, a more universal tool is being developed, designed not to solve any specific problem, and a user-oriented product is being created. Currently, there is a trend: scientists are interested in working with neural networks, but are not ready to use them if they are not applications, — Anna Nartova explained.
Anna Nartova noticed that an important part of the congress was the section on chemistry. This is due to the fact that the Chemistry Congress was combined with the Assembly of the Jupak, so not only purely scientific issues were considered. It is very important that it was, first of all, about teaching schoolchildren chemistry. Teachers in this subject became a problem for many countries whose representatives participated in Congress. In the profession of “teacher”, and in particular the “teacher of chemistry”, they go reluctantly: it is labor -intensive and responsible, but, unfortunately, does not belong to the highly paid. Colleagues from different countries noted, based on their own experience, that in order to attract students to this science, it is necessary to focus not on high -grade, but on students of the middle or even junior school link. Section participants shared work experience in different countries. The Indian chemist Udi Maitra made a report “Improving the perception of science/chemistry through simple experiments associated with everyday life.” This project is of deep humanitarian significance, spreading the natural science worldview, as well as increasing the motivation for education. The scientist from Taiwan Khsien Chang Cao presented the message “Chemistry on wheels”, in which he spoke of a huge personnel hunger, growing since 2017. Working on the problem of attracting personnel, scientists -enthusiasts with mini -laboratories go to provincial schools and show the children experiments with various chemicals and in exciting form tell the children about chemistry as an interesting science. Many of these guys at the end of school make a choice in her favor and return to their schools already as teachers of chemistry or become volunteers of educational projects. It is interesting that this work is supported by both the state and corporations, since they are very interested in improving the level of school education and attracting motivated teachers to the school. Anna Nartova noted that this work exactly corresponds to the concept of the project “Entertaining science for schoolchildren”, which has been implemented at NSU since 2009, and in recent years has been part of the school center of the Faculty of Physics of NSU. And we had something to share with colleagues in this area.
Material prepared by: Elena Panfilo, NSU press service
Please note: This information is raw content obtained directly from the source of the information. It is an accurate report of what the source claims and does not necessarily reflect the position of MIL-OSI or its clients.
.
