One of the World Bank's steps towards ending extreme poverty is to reduce global poverty below 3% by 2030. At the national level, 193 countries adopted the Sustainable Development Goals in the United Nations Development Agenda to 2030 on 25 September 2015.
Russia sets an even more ambitious goal – to halve the poverty rate in the Russian Federation by 2024.
In order to achieve our goals, we would like to obtain solutions using AI&BigData in the following areas: 1) Creation of the most accurate algorithm for rapid search and identification of the poor - accurate measurement of the indicator itself with the maximum decomposition of the characteristics of the poor; 2) Determination of the true causes of poverty based on open data from different sources; 3) Comparing the poverty programs of countries and regions with the dynamics of real indicators and determining the effectiveness of support measures/programs/strategies - monitoring the effectiveness of poverty reduction measures. 4) Given the current situation in the country and the world, the issue of forecasting the number of poor people is also acute. It is necessary to build a model that will predict the number of poor households in the region as a whole - a period of 1 year, 2 years, 3 years, monthly.
Tasks and Data
Qualitative change in the interaction between state and an individual
To assess the regions a set of indicators is planned to be formed in the following areas (during the development of the methodology areas can be adjusted):
Housing, communal infrastructure and communications
Sports and leisure
The contestants are invited to propose a list of indicators in 3 or more areas, which should be included in the rating, collected on the basis of data with the help of algorithms in a regional and municipal regard, as well as to present a list of possible management conclusions that can be made based on the findings.
Tasks and Data
Rapid changes in economic sectors and technology could cause significant disruptions to labour markets in the coming years. Businesses around the world may face risks due to a shortage of qualified personnel. As a result, universities that do not have a clear understanding of market trends may offer training in an irrelevant volume and quality, and people may choose professions and retraining courses that are already outdated.
Governments must accurately assess future prospects on the job market when updating their education policies. It is necessary to systematically forecast the evolution of skills requirements in order to ensure the development of relevant training.
The challenge is to use artificial intelligence technology and data analytics to solve problems in the following areas:
Creation of algorithms for predicting "future skills" (skill sets relevant to future markets) based on available open data (including kaggle, Google dataset search, job search, HeadHunter, Linkedin, etc.):
Algorithms predicting future technologies on the basis of scanned scientific articles and patents that identify emerging technological trends.
Algorithms predicting introduction of "breakthrough" technologies. Prediction of technological evolution in existing or emerging market areas.
Algorithms predicting " hype" technologies (with a high probability of becoming "another major phenomenon", temporarily overheated by the need for specific technologies and skills).
We refer to the "skills of the future" as:
Skills associated with "disruptive technologies" (rapidly changing, new and emerging skills associated with such technologies and tools).
Skills development (skills that can change over 5-7 years).
Literacy emergence (the sum total of necessary knowledge and skills associated with a particular interaction or activity).
Soft skills (focusing on the most important soft skills).
Trend-setting skills on a hype.
Development of diagnostic tools for identifying skills in the above mentioned technology stacks. This model should be useful for:
skills that require intensive professional development or retrainin
Automation in identifying training and skills development resources that are relevant to the needs of organizations and applicants. Automatic recognition of digital texts, audio and video.
Tasks and Data
Braille alphabet recognition
The task is to automatically recognize texts written in Braille and convert them to Cyrillic for Russian and Latin for English. Recognition is made from photos and scans taken without the use of professional equipment. The solution of this problem is aimed at facilitating the social integration of blind and visually impaired people and expanding their communication opportunities.
Over the past few years there have been more and more devices designed to improve the quality of life for the blind and visually impaired: voice assistants, "Braille line", etc. These tools certainly help in communication between the blind and the sighted, but we want to take another step "from the other side" - to help sighted people read the text written in Braille.
Sighted people usually can't read a Braille text, and they have to make a lot of effort to work with white dots on white paper. It is especially difficult to read texts when they are printed on two-sided braille, when the convex dots alternate with the recessed dots on the other side of the sheet.
Relatives of the blind and visually impaired, for example, face the problem of checking a child's homework due to lack of Braille reading skills.
At present, there are no compact and inexpensive solutions to automatically translate text written in Braille into Cyrillic or Latin. For texts in Russian, there are only cumbersome and expensive scanning machines with built-in Braille recognition. Price and size make the use of such machines in everyday life absolutely impossible.
The task is relevant, because a system of automatic recognition of texts written in braille will:
help to attract teachers who do not know the Braille alphabet to teaching blind and visually impaired;
make it easier for teachers who work with the blind and visually impaired in inclusive programmes in ordinary classrooms;
will allow reissue of Braille textbooks for which there is no digital source;
will allow blind people to submit written requests to various authorities themselves when needed;
generally facilitate communication between sighted and blind people.
The importance of solving this problem quickly is due to the participation of blind children in the All-Russian Olympiad of Schoolchildren. Currently, works written in Braille are first deciphered by hand, and then checked by the jury of the Olympiad.
Tasks and Data
Development of a training simulator based on an environmental weather forecasting system
Currently, the Ostankino TV Tower has a meteorological complex collecting climatic and environmental data.
Measuring radius: Moscow.
The complex consists of twenty-one suspended meteorological booms located at different heights, as well as one ground meteorological sensor. Devices record temperature, air humidity, cloudiness, wind direction and speed, pressure, thunderstorm activity, etc. The uniqueness of the data is that they are obtained from the highest located meteorological equipment in Russia.
Among other things, on three levels there are devices measuring atmospheric pollution parameters: nitrogen, carbon dioxide, dust, etc.
Based on a set of environmental and meteorological data over several years, it is necessary to develop an environmental meteorological forecast system, which establishes a pattern between changes in weather data, atmospheric processes and the environmental situation in the city. As a result, this system will provide more relevant weather information for different audiences. The developed system should also be used to create a training simulator for schoolchildren, students and other scientific communities to demonstrate climate change depending on environmental and atmospheric factors for both theoretical and practical simulations (weather forecasting based on real current climate data and theoretical forecasting based on random variables).
Tasks and Data
Analysis of key barriers for people with disabilities (based on open data)
According to statistics, about 12 million people with disabilities live in Russia, from 40 to 80 million people with health limitations. In total there are more than 1.3 billion people with health limitations in the world.
At the moment there is no comprehensive study that provides insight into the barriers and needs of these people in various service delivery areas.
There is a need to analyse open sources:
customer testimonials on websites and forums;
specialized groups and forums of people with disabilities, etc.
Nosologies that should be considered:
the visually impaired and the blind;
the hearing impaired and the deaf;
the blind and deaf;
low mobility groups (wheelchair users with artificial limbs, elderly);
people with mental and developmental disabilities (Down syndrome, cerebral palsy, autism spectrum disorders, dementia, Alzheimer's disease, etc.).
Areas of service:
leisure and entertainment;
Often, businesses do not consider people with special needs as their potential customers, so many services and products remain inaccessible/unsuitable for people with disabilities and health limitations.
There is a need for a rationale to enable companies in the B2C sector to adapt services and products to different customer categories.
There are statistical data on gender and age, numbers by region, by disability group. But there is no comprehensive market research that focuses on the preferences of people with disabilities, difficulties faced by them, and their purchasing power.
Service to detect the development of cardiovascular disease
Creation of an algorithm for assessing the probability of vascular catastrophe. Creation of a digital patient profile within the personal office of the Unified State Information System in the sphere of health care (USISH). Timely notification of the patient by means of automatic mailing of an invitation for CVD prevention check based on the results of Machine Learning. Notification of the physician through the patient risk group service in the USISH and development of individual recommendations for examination and treatment.
We hypothesize that the relationship between physical health data (evaluation of cardiovascular risk indicators), work activity and the fact that a CVD has occurred makes it possible to create an algorithm for assessing the probability of vascular catastrophe (heart attack or stroke). Specifically, it is possible to create an algorithm for assessing the probability of vascular catastrophe (heart attack or stroke):
blood pressure level
the fact of smoking
place of work
the fact that there is a CVD.
We propose using the European scale for the calculation of the risk of death from cardiovascular disease in the next 10 years "SCORE" as a basis for the algorithm assessing the probability of vascular catastrophe.
Tasks and Data
Verify the accessibility of digital products for people with disabilities
The goal is to develop a system that allows you to assess the level of accessibility of a website or application for people with various disabilities: total and partial vision loss, vision impairments, total and partial hearing loss, total or partial loss of speech functions, physical limitations when using a digital product, mental disorders.
Accessibility is assessed in accordance with the following standards of The World Wide Web Consortium (W3C), GOSTs and recommendations of the Russian government entities.
The system should be in Russian and English, providing not only information about errors, but also recommendations for their correction.
The system should include an assessment tool with a clear and simple user interface, allowing to evaluate both exsiting sites and applications, and prototypes.
The system should show the level of accessibility on a predefined scale (e.g., percentage or levels) that will be set by the developer in general and for each individual nosology in particular (to what extent it is accessible to people with sight, hearing, speech, physical and mental disabilities), indicate errors and suggest corrections, provide a detailed report on the level of adaptation of the site/application's individual functions. The report should be fit to be sent by mail and to be viewed online. The report should be understandable not only to programmers, developers and designers, but also to ordinary users who would like to know if the site is accessible or not. Based on the results of all checks, the system suggests an accessibility rank of verified sites, defining them in different accessibility categories (e.g., green, yellow, red). The rank should be hidden and available only to the platform administrator.
If possible, it is preferable to design a system that independently factors in new requirements for sites and applications, because W3C, Apple, Android are constantly improving their requirements, and changes must be taken into account.
Tasks and Data
Assistant professor, Inha University in Tashkent (Uzbekistan)
Professor, Almaty Management University (Kazakhstan)
CEO of Ecosystm (Singapore)
COO of Ecosystm (Singapore)
Head of Partner Acquisitions at RecruiterPal (Singapore)
Senior Operations Officer, Data Analytics and Tools, The World Bank Group (USA)
Founder of Multiverz (Singapore)
Lead Economist, Poverty Global Practice The World Bank (USA)
Rachel Alexandra Halsema
IT Officer, Information and Technology Solutions, The World Bank Group (USA)
Head of the ASI Digital development centre
Head of program of the ASI Digital development centre
Collecting data on poverty today is time-consuming and is often delayed. Besides, we do not always have accurate data. And most importantly, we do not understand the key reasons and causes of poverty
As a result, decisions are often made based on old data, without analyzing their nature. Machine learning can dramatically change this situation, making poverty measurement more accurate and much closer to real time analysis, helping to find key reasons for its occurrence. This will accelerate ending the poverty in the world
We would like to explore solutions based on AI and data analytics in the following areas:
1. Creating the most accurate algorithm for a quick search and identification of the poor - an accurate measurement of the number of poor with the maximum decomposition of their characteristics.
2. Determining the true causes of poverty based on open data from various sources.
3. Comparison of countries' poverty alleviation programs taking into account the dynamics of indicators, and determining the effectiveness of support measures / programs / strategies by monitoring the effects of measures aimed at the elimination of poverty.
When solving a problem, it is necessary to consider the international experience. Solutions must be prepared for one or several countries involved in the project (Russia, Singapore and other participating countries), with a breakdown into regions and cities.
Tasks and Data
Остались вопросы? Свяжитесь с нами!
Нажимая на кнопку, вы даете согласие на обработку персональных данных