it is open source principles thanks to the work of many thousands of programmers who create open libraries of solutions and developments, and therefore , in the field of introducing artificial intelligence, we will definitely strengthen cooperation with our interested partners, as i have already said with everyone, but including brix and conduct joint research and educational programs to maximize the potential of advanced technologies for the benefit of our countries, some relevant cooperation documents will already be signed and continue to work in this direction, dear friends. what results we will achieve in the practical implementation of artificial intelligence technology depends on our joint efforts. from the readiness of businesses to increase their technological and improve their technological position from the desire of the authorities on the other hand to increase speed. the pace of change will pave the way

for the introduction of advanced technologies starting from next year 2023. we will monitor the results of the use of artificial intelligence by sectors of the economy and social sphere. to do this, i propose to create a special tool. the index of maturity of the regions' branches, but so that it can be more sonorously said the index of intellectual maturity, because we are talking about implementation. e, means artificial intelligence and, on its basis, practically evaluate the results of the work of each subject of the federation.

in this regard, i propose to establish an annual award for young scientists and engineers for scientific and design breakthroughs in the field of artificial intelligence. and one more award for success in the field of artificial intelligence implementation. also annually hand over to domestic companies to enterprises in regions and cities. and the solemn awards ceremony, e, for the first time could be held just e in the next year 23 and at the next conference, the journey of the world of artificial intelligence. i thank you for your attention, for your patience and sincerely wish you success we kind of agreed with the german that my work will end there today, uh, due to the workload, but uh, knowing that i looked. uh, what kind of ideas do colleagues have, uh, in

the course of our conversation during the demonstration of various projects, so i propose and think that it will be right. if, after all, i listen to the proposals that may be made, then you will continue in the course of our joint work. e additional express your wishes on your proposals. colleagues we will fix everything, we will certainly process it all and, uh, we will try to implement it. thanks big once again for your attention. chose three short reports and three speakers of young scientists who will present their proposals here. these are leaders in today's relevant applied areas for the development of the implementation

of artificial intelligence. well, the first topic is probably such a topic, uh, that today all, uh, very much all researchers. uh, they are very stimulating and it is connected not only with, uh, increasing the efficiency of the industry, but the most important thing is with saving people's lives, and you said this about health. this is our health care and therefore. uh, we want to start with this particular topic, we, uh, also uh, consider this topic a priority. we have a special company called sbermed , and with our partners, first of all, of course, with moscow, we are introducing new modern solutions, and then these solutions are being implemented in the partners of our regional 63 regions

today. these are our partners who have implemented e, a large stack of artificial intelligence solutions and continue to further develop this technology and let us provide e first scientist elena vladimirovna sokolova, who will talk about what she does in the field of healthcare. elena sokolova began her career in pharmacology, where she created new biological products. i am the kind of person that i want to see results fairly quickly in the short term. i do swimming. i really love to sing her interest in computational biology led her to the artificial intelligence laboratory. sberbank. whether it is possible to reveal an early stage of a diabetes by results of the electrocardiogram. i'm proud to be we were the first in russia to develop and implement a solution for analyzing medical texts in order

to help doctors make preliminary diagnoses. we made this task from scratch and we managed to implement it. in all adult polyclinics in moscow, i count the point. yes, the services that elena creates today diagnose the disease according to the electronic card with a medical image and signals and have never liked it. uh, you know the way to go with the flow is within. that's smart, brave and beautiful. have a good afternoon colleagues, and i i represent the sberbank artificial intelligence laboratory today. i want to talk about how we are already solving pressing problems in healthcare with the help of artificial intelligence. today , the two leading causes of death, both in the world and in russia, remain cardiovascular and oncological diseases. at the same time, in 2019,

cardiovascular diseases caused almost half of all deaths in russia, and due to late detection of oncology , 120,000 patients die within the first year. after the diagnosis is established. that is why our solution is aimed at the early detection of these diseases. the main areas of our work are focused on the analysis of medical images and the analysis of medical texts and now i will tell you about an example of one of these projects that we implemented in the laboratory of artificial intelligence for analysis, and ct of the chest and which was introduced into practical healthcare with the help of one of our , and the companies of our ecosystem are sbermet and during the pandemic, you all know what was done simply a huge number of ct scans of the chest. it was a very overburdened public health system and while the attention in the calculation of radiologists was directed to the assessment of viral pneumonia. we trained

artificial intelligence. on these complex data, find signs of lung cancer. in one of the above experiments in a real clinic , 1,500 ct scans of the chest were analyzed on real clinical data. of these, artificial intelligence identified 12 patients with a high probability of lung cancer in eight of them the diagnosis was confirmed and the treatment of these patients has already started. patients. but compared to medical images, the largest amount of data about the patient is accumulated in the form of texts of poorly structured, unstructured and complex channels, these are complaints, and we have this and this instrumental laboratory analysis, and so on. in fact, the greatest potential for the health system lies precisely in the field of analysis of such data of data texts. now in medicine there are two main trends. the first of

these is a population analysis of the population and the identification of people at risk. development of complex chronic diseases. for their timely prevention. the second is a personalized analysis of a huge amount of data about a particular patient in order to develop an individual approach, and to provide him with medical care. artificial intelligence is able to process any, uh absolutely yes, volumes of information in a short time, the doctor will not physically be able to study the entire medical history. about each patient that is why i believe that an artificial intelligence service should become an ordinary assistant to every doctor. one such example diagnostic assistant this model, which we are developing together with the government of moscow, this model is trained to an unprecedented extent on a large amount of data. this is about 30 million visits of patients to moscow over 2 years, this model is analyzed, and all

information is offered by the medical ops to the doctor, and the final diagnosis. as a second opinion. this is an absolutely unique complex for in-depth diagnostics and monitoring of the patient's condition, which has no analogues either in russia or in the world. and what is important to note this model, we trained with moscow spins. these are specialists of the highest qualification, respectively, by implementing it in the regions, we can provide an affordable high level of diagnostics throughout the country. these are just small examples of what we have been able to do as examples. actually a lot of time. it's not enough to tell everything. e. well, what i would like to end this with, and what, in principle, has already been said, the house of his speech. these are the barriers that need to be overcome in the industry for large-scale and accelerated implementation, and artificial intelligence models of the workshop and the first such barrier is legal restrictions.

accordingly, we ask to simplify, and legal regulation in terms of access to impersonal medical data, and for learning to test the model, because in the data of sans in building artificial intelligence models, of course, there will be no data, but nothing as a result, the second restriction or even not restrictions, probably, very serious steps have already been taken in this direction. this is the implementation and evaluation of the effects of the artificial intelligence model in real practice and here, and again a unique very successful project - this is a moscow experiment. in europe, about 7 million oe of radiodiagnosis images have already been analyzed , which is approximately 40% of the total volume of radiodiagnosis research in moscow, and the creation of such platforms should become a guarantor of the uniform quality of such services and their safety. this in turn will help us

overcome some skepticism from the medical community. of course they should understand. what is the quality and how safe our decisions can only be made through their implementation and clinic economic evaluation, and the effects after implementation. and here i would like to say that, in our opinion, such sites are necessary, but they should be scaled up, and moved outside of moscow, scaled up to other regions, and we believe that we can start from the moscow region, both in terms of technological maturity and in terms of view of law-regulatory maturity. this is a fairly well developed region. and the third after successful implementation and evaluation, a such services in practice, they need to be included in clinical guidelines and in the standards of medical care, i conclude my speech, i want to say that i sincerely believe that the power of

synergy in any industry, as well as in our branch of medicine, synergy of artificial intelligence and the medical community will allow us to achieve . a common in my opinion, our goal is to extend the average life expectancy up to 78 years, and it will be better up to 80, and within the coming decade. i believe that the introduction of artificial services intellect. it will definitely help us to achieve it and even faster than we think on this. everything. thank you so much. i said that i would like to listen to you and mark some things for myself and then sort of analyze them, but i cannot help but react, nevertheless, to what you were talking about. uh, health care is the most important activity of any state, but the duration and quality of human life. this is such an aggregated generalized indicator. e work of the state in all areas, both in

the economy and in the social sphere. uh, and definitely. uh, joining the 80 plus uh club is a task that remains with us unconditionally. uh, to this we will move to this we will go. eh, and we will also do this with the help of artificial intelligence tools, positive examples already exist, and not obvious. absolutely. we will certainly do all this. we will remove obstacles to the use of these tools. i spoke about this in my speech. eliminate barriers need to analyze all this. i will be grateful to you. if you complete all this in an appropriate way on and. uh, authorities. together with you , all these three components that we have just mentioned, including feedback and expansion, will make the appropriate decisions. eh, we will certainly implement this practice. i already mentioned. eh, about the

moscow region. now they also said about it. here, i would ask the governor, uh, to say two words. yes, thank you, dear vladimir vladimirovich we also try to be friends with artificial intelligence. we are implementing it in various areas. ah. we have everything for this required digital information. you were with us in the control center, region and amount of data. for each direction, it is enough to optimize a here and look for an effective solution, but in particular in medicine. we have n't read the pictures yet. we will have it next year, but, uh, we have a robot svetlana which, accordingly, receives 35,000 calls every day and , accordingly, helps people get to the clinic or call a doctor at home, and the key task is to further develop, and the artificial

intelligence, and so that it is already a function, and you once supported, and in the decision to control the disposal of construction and household waste from this year. ah, if there were before, two slides, if possible, yes, yes, that's how we had a couple of years ago, the environmental supervision inspector had to stop the car in his hand, write out, uh, a resolution and fine if there was garbage unauthorized after a legislative decision, respectively, already a video camera and artificial intelligence, but issues a violation, because the electronic coupon for each car is tied, well, until now, the inspector. e, accordingly, writes out this decree, and next year, fully artificial intelligence will be, e, the functions will replace the inspector in this way. and we will have

a fairly noticeable progress in this area, and it is obvious that we need a transformation, for this we need partners. we are very pleased that we have the opportunity with the asber team. work in all these areas. naturally, what about moscow? uh spoke speaker e about synergy. i agree, that, including in partnership, but a lot of very significant effect. and i really hope that next year we will show noticeable noticeable results, of course, in conclusion, vladimirovich needs to be prepared for the fact that we will apply for support in changing legislative regulations, because they are often an obstacle to the further introduction of artificial intelligence and everything new, that's, uh, in general, what i wanted thank you, regarding access, uh, easier access to medical data, i also agree that there are features,

of course, uh, when it comes to data, and depersonalized, especially since we will push the adoption of these decisions to us. thank you good luck please. thank you so much. thank you elena thank you. to ask evgeny vladimirovich or who is the director not almazov, who is one of the leading not only medical, but innovative institutions that introduce digital technologies and artificial intelligence. vladimir vladimirovich, we would like to, dear colleagues, a wonderful project that fully corresponds to the trend that is healthcare. i would even say more probably, it belongs to the area where artificial intelligence is definitely already working, in some areas better than specialists in medicine, and even a question arises. will artificial intelligence be replaced by specialists? yes, of course, it definitely won’t replace it, but what will happen in the future is probably a specialist who

owns artificial intelligence, but it will be better to treat patients and ultimately replace the specialist who does not own artificial intelligence. it must be said that the performances that were presented by elena vladimirovna of course, they reflect that progress and the depth of penetration of artificial intelligence in health care. and if you allow during the speech. i would like to focus on three aspects that arise while listening to this project, which is very interesting and innovative. eh, the first direction is all the same. eh, we should think about how to stimulate the development of medical technologies on the basis of the institute, how to implement them today, a lot is being done in the country and e. a vertically integrated medical system is being created a single health care loop the vast majority of regions are already participating in these projects , transmitting structured data to a productive

environment, but if you look at the real situation of only 20 e today, medical, and technologies using artificial intelligence have registered an increase in equal supervision, here, of course , there are questions related to the regulatory. what did you talk about or one question related to promoting the participation of business, and uh, we would probably really ask for a war, if possible, not only a clinical recommendations, but in the mandatory health insurance system, those technologies that are created on the basis of artificial intelligence should be treated as resource-intensive and evaluated differently. and thus it would probably be right to encourage companies to create such technologies, and the federal project on artificial intelligence should provide for . e, a means or event related to the stimulation of the creation of data sets that can be placed for general use and many

companies do not spend extra money on quite laborious work. could create artificial intelligence products on this basis is the second question. what you didn't want to touch. this is a question related to, uh, those solutions that will really change the organization of medical care itself. today we are moving towards personalized medicine. this is largely due to artificial intelligence and largely due to digital solutions. and this is for sure, because the development of genetics and personalized medicine. impossible without bioinformatics, without artificial. ah, artificial intelligence. and in this part. of course, we need integration solutions that can unite all stages of medical care. make it seamless. her e, i would say, is focused on the end result and the patient is oriented here. e, of course, we must think about some solutions, of course, there must be

platform solutions. what you already said in your speech and uh, we are very supportive. this direction is precisely a platform solution, namely the transition from reactive proactive control to based on big data, and using not only those digital solutions that are available in every institution, but most institutions. today we can call digital clinics. here is the transition to such e to analytics, but to the decision support system and the system. e. i would say, not so much predictive, but priscriptinum management, this is what today characterizes modern trends and characterizes the state of the clinic, from which one can call value management. here, of course, the use of e is very important. in the current activity institutions and all data that are associated with patients, especially high-risk patients with a

high risk of adverse events. here , artificial intelligence can accurately help medical personnel identify these patients, predict the development of adverse events and help. uh, help avoid them. i must say that, of course, today we are developing a lot in this direction with the ministry of health and sberbank. uh, these are questions related to speech technologies. what we talked about today. these are issues related to the creation namely the smart clinic and, uh, those technologists. the solution that we plan to get in the near future can indeed be scaled to the level of the whole country and the third section. hey, what were you talking about too? is it education? of course we need programs for students. we need programs for doctors, but today we are within the framework of the specialty project that you allowed us to train

in a scientific institution, and we are starting to form an individual trajectory from the fourth year student education that is based on the use of artificial intelligence to analyze the preferences of each student, his personal experience, his uh, practical skills that he has and uh, those scientific achievements that he already has in his senior years. it seems to me that it is also very important, of course, very much. obstacles to difficulties for the introduction of artificial intelligence in health care - this is organizational - this is technical and so today we see this is artificial force, artificial intelligence, which you will help us, of course, working with raw data is an ethical issue, which you talked about the protection of personal data and i think that in a joint effort, of course we must move

in this direction, these funds that will be invested. they are definitely trying to dig in in the form of reducing the incidence of mortality and saving the lives of the citizens of the russian federation, of course, the fulfillment of all the tasks of the tasks of the federal project that you have set. thank you for your attention. she is not yet know that you have, uh, many modern technologies are being actively implemented. uh, heard and noted about uh, creating an incentive system. e, the development of these technologies, including artificial intelligence, and in clinical guidelines and in the system a, it is mandatory for medical e-e in the whirlpool system. so we're going to do all this. e, about integrated solutions , also noted the use of hmm platform solutions. uh, i don’t know, we have mikhail albertovich 10 no, well, we’ll talk to him or in the near future he, uh, a modern person knows about all this and

the ministry of health is also trying to be in trend and doing many useful things, but for this purpose today we have gathered herman and gathered everyone in order to see what needs to be done in the near future in various areas, including in syria , we will work with artificial intelligence . thank you. big. thank you. it seems that he also wanted to give an opportunity to speak to a very bright and learned young man, who is already a laureate of the president's prize. uh, among young scientists and ivan the master is a researcher. mathematician from call-tech ivan if possible. e a few words about what you do in terms of. e creation of a base for the development of artificial intelligence and optimization of models, firstly, vladimir vladimirovich thank you for the excellent praise. it's nice. i'll try to briefly say what we do, so that it is clear what exactly praise a. in fact , now the development of artificial intelligence

comes from a large number of small private models, and so weak artificial intelligence to large models that are trained on dozens hundreds of billions of examples, they have hundreds of billions of parameters, but they are able to solve a large number of different cognitive tasks for training such models. supercomputers are used, for example, the christofari supercomputer , and the power consumption of the christophoris supercomputer corresponds to the power consumption of about a city with 10,000 inhabitants. but even on such a supercomputer, training the kandinsky model took 4 months. and we can also say that the total capacity, and of all the data centers of the savings, is approximately the energy consumption of the city of prominent large imagine a whole city of computers that consume electricity, that is, 4 months. this is a lot. and why is this happening? well, we estimate the current efficiency of open source model training software at about 30%, we want to roughly

double this figure. what, uh, what are we doing in this direction a year ago, within the framework of the federal project for supporting artificial intelligence centers, second-hand ministry of economic development and savings, we launched a green artificial intelligence project, just one of whose goals is to increase the energy efficiency of training the model , the first result was already obtained in the shortest possible time. we reduced the memory temporal training of the model by about 15%. this actually translates into tens of hundreds of billions of rubles, as you did when training the model, we use hundreds of billions of parameters that need to be updated and one number that we store takes about 32 bits. this is for understanding. just seven decimal places. well, it would seem. why such precision? yes you can store two characters and everything will be fine. but it turns out that keeping only two quality marks of the final model will seriously fall, which we did using mathematics, using a fundamental mathematical approach. we found those places in the process of a complex learning process, where we can replace 32 bits with three memory in this

place, reduce by eight times, and the total time by 15%. clear, in fact, as i said, our goal is still not 15% a twice, and therefore we started development. actually domestic software, based, in fact, on fundamentally different principles. well, i’ll also give such, probably, an understandable example, suppose. we have a big family, a lot of shoes, big boots, small children's shoes. they need to be sorted into boxes. all the same, there will be empty space in the boxes. and this empty space corresponds to the inefficiency of using the computing system, how we solve this program very simply by taking computing blocks. cut into pieces, shoes. we cannot cut into pieces, but computational blocks can be cut into pieces in this way. uh, we already have preliminary ones. results. we expect that we will get our own software and the efficiency will really increase, but instead of 4 months there will be two, we still have enough time to train one model. to do this, you need to increase computing power the standard way to increase computing power is to buy more computers, but they are all based on microprocessor technology, and it has already

come up. the size of one transistor is about 10 sizes of atoms, that is, yes, 5 years and further increase in this area is completely impossible, so state companies around the world are looking in the direction. so that is alternatively some word for not very good looking computing systems. actually in order to understand what is a promising computing system. it can be, in fact , a completely arbitrary physical system. well , let's fantasize about it. we have an algae aquarium. we shine with light, a chemical reaction occurs as a result of chemical reactions. we read the result, that is, the task of a modern scientist is to actually develop, pick up such a physical system and write a program for it. naturally, one of the most promising areas. here here is shown the famous schrödinger's cat is quantum computing, which is built on the principle of superposition. we don't know if the input is alive or not until we open the box, but a quantum system that requires 50 quantum bits on a classical computer.