A few days ago, the "Artificial Intelligence Data Science Competition White Paper 2022" was officially released on May 19.
"White Paper 2022" is based on the situation of domestic competitions in the past three years from 2020 to 2022, focusing on the differentiated development of various competitions by governments, enterprises, and scientific research institutions, dismantling the basic structure of competitions, analyzing the development difficulties of competition formats, and studying and judging artificial intelligence and data science competitions Future trends and development directions.
"White Paper 2022" focuses on the current situation of artificial intelligence and data science competitions in the past three years, actively explores the value path and development direction of artificial intelligence and data science competitions through sampling surveys, interviews, sample analysis and other research methods, and provides digital transformation for many enterprises and institutions Bottlenecks provide preliminary foresight, and provide innovative ideas for expanding data application methods, promoting the circulation and application of data elements, and finding excellent data talents.
( The full text has a total of 39211 words, this article is an overview, please check the full HD file at the end of the article)
Overview and basic value of artificial intelligence and data science competitions
Overview:
The artificial intelligence and data science competition is a data application model in the form of competition. It gathers a large number of digital talents in a short period of time, and promotes the development of artificial intelligence applications and the value mining of data elements through reasonable competition design.
Basic value:
As an innovative format, the artificial intelligence and data science competition brings together resources from industry, academia, research and application, and influences and helps the entire industrial ecology externally. In the data science ecosystem, the government, enterprises, technical talents, and competition organizations form a subsystem of a virtuous circle, which innovates the circulation mechanism of data elements and talent elements in the system, and empowers the development of the data science industry.
Development and Trend Insights
Differentiated development of various competitions
government events
Focusing on public utilities and industry applications, digital twins empower the construction of smart cities
In recent years, the proportion of special events in the data science event market has gradually increased, but in government-run events, comprehensive events are still the mainstream. At the same time, the development of the digital economy and the construction of smart cities have become the key tasks of government departments in many places. Artificial intelligence technology represented by digital twins plays an active role in the construction of smart cities.
Corporate events
Explore the deep integration of competitions and communities to build an active user ecosystem
As an important part of the technology community, artificial intelligence and data science competitions are used by large technology companies as an important means of community drainage and user retention. At the same time, through competitions, community user participation is enhanced, user stickiness and community technology concentration are increased.
Scientific research competitions
Independence is weakened, try to bind government and enterprise events
The artificial intelligence and data science competition was born in academic conferences, and existed as an academic research activity for a long time at the beginning of its appearance. However, in recent years, the proportion of academic institutions running competitions independently has continued to decline, and cooperation with the government and enterprises has gradually become the main competition for scientific research and academic institutions. choose.
Question Trend
Computer vision (CV) competition questions are the most popular in 2022. Among the 635 competition questions counted in 2022, computer vision (hereinafter referred to as CV) related competition questions accounted for 36.5%, reaching 232, which is the technical direction with the largest proportion .
The application of focus on industrial events
Since 2021, the number of industrial AI and data science competitions has increased year by year. In the 12 industrial competitions in 2022, all the algorithm competition questions are designed around practical problems in industrial production. The difficulty of the competition questions is moderate, and the competition plan is easy to implement.
With the popularization and application of industrial Internet and the trend of digital transformation of traditional industrial manufacturing, industrial manufacturing enterprises have begun to pay attention to the comprehensive and in-depth perception of industrial data, real-time transmission and exchange, fast calculation and processing, and advanced modeling and analysis to realize intelligent control, operation optimization and production Organizational changes.
Competition subject
Sponsor
The sponsors of AI and data science competitions can be divided into three categories: governments, enterprises and scientific research institutions.
Among the 211 competitions counted in 2022, corporate competitions are the mainstream, with a total of 109 competitions, accounting for 51.7%; the number of government competitions is more than that of scientific research institutions, with a total of 64 competitions, accounting for 30.3%; scientific research institutions (including Universities) held 38 games, accounting for 18%.
government
Although government-sponsored competitions only account for 37% of the total number of AI and data science competitions in 2022, they have a greater influence in the overall competition ecology.
Analysis of the current situation and trend forecast of government competitions:
-
Explore the application of public data and promote the incubation of innovative projects
-
The level of competition remains high and the competition is attractive
-
Government data is the mainstay, and various types of data are assisted
-
Able to provide high-quality scarce data, high-quality technical proof, and highly attractive to high-level teams
enterprise
Data-driven is the core way for enterprises to achieve digital transformation. Artificial intelligence and data science competitions can make up for the lack of manpower, time, and cost in the data application link of enterprises in the process of digital transformation. Some competition questions go deep into the data processing link to explore complex A new approach to data processing.
Analysis of the current situation and trend forecast of enterprises running competitions:
-
Running the competition takes both talent reserve and brand building into consideration
-
The direction of the competition is diversified, and the data-intensive field is still the mainstream
-
Artificial intelligence and data science competitions become a new path for corporate philanthropy
Research institutions
Scientific research institutions focus on the output of scientific research results and the cultivation of scientific research talents. Its competition focuses on precision, and attracting top talents in the field to participate is the focus of its competition.
Analysis of the current situation and trend forecast of enterprises running competitions:
-
The commercial atmosphere of competitions sponsored by scientific research institutions and universities is relatively weak, and they mainly focus on two aspects: disciplinary competitions and interdisciplinary scientific research.
Competition platform
The competition design capabilities, technical support capabilities, and competition operation and publicity capabilities required by artificial intelligence and data science competitions may exceed the capabilities or responsibilities of some sponsors. Therefore, the event platform has become an important third party that transforms data resources into a complete competition.
The main responsibilities of the event platform include three aspects: event design, technical support and event operation.
Third-party independent platform
dc contest
The DC competition (DataCastle) platform was officially launched in 2016. It was initiated by Professor Zhou Tao, director of the Big Data Research Center of the University of Electronic Science and Technology of China. Office, computing resources and other services.
Based on the Kaggle competition model, the DC competition adopts a platform-based, modularized, and automated way of running the competition. At the same time, it combines the specific needs of running the competition in China, based on the self-developed data science training platform DCLab and related patents, to provide customized running for the organizers . Race service. After years of competition practice, the DC competition platform has developed into a leading third-party competition service provider in China, with more than 325,000 registered users, more than 500 online competition questions, and a cumulative bonus of more than 97 million yuan.
and Whale Community
Hejing Community (formerly "Kesai Network") was established in 2015. It is one of the well-known third-party data science communities in China. It is an earlier batch of platforms that focus on big data algorithm competitions. It has nearly 200,000 registered data scientist users. Radiating more than 300,000 data talent groups.
DF contest
DF Contest (DataFountain) is a brand of Beijing Shulian Zhongchuang Technology Co., Ltd. It is a leading data competition service platform and data intelligence collaborative innovation platform in China, aiming to form a professional growth chain for big data enthusiasts around collaboration, data, knowledge and skills Road, empowering data scientists and industries.
Enterprise self-built platform
Tianchi
Alibaba Group officially launched the "Tianchi" big data scientific research platform in 2014. Based on Alibaba Cloud's open data processing service ODPS, the platform opens massive data (Alibaba data and third-party data) and distributed computing resources to the academic community. The platform business includes : Tianchi big data competition, data laboratory, open teaching, data talent certification.
Paddle AI Studio
Flying Paddle AI Studio is an artificial intelligence learning and training community based on Baidu's deep learning platform Flying Paddle. It provides an online programming environment, free GPU computing power, massive open source algorithms and open data to help developers quickly create and deploy models. There are different competition categories such as paddle flying competition, regular paddle flying competition, and rookie practice competition.
Huawei Cloud
The HUAWEI CLOUD Contest is a comprehensive competition platform for developers created by the HUAWEI CLOUD developer platform. The competition covers machine learning, software development, hardware development, system development, industrial Internet and many other fields. Strictly speaking, the Huawei Cloud competition platform is not a data science competition platform, but a comprehensive developer competition platform, and it only serves Huawei Group's own business.
contestant
Identity
Educational characteristics: In terms of educational background, about 60% of the participants are masters and doctoral talents, and undergraduate talents account for 37.88%. In terms of talents and professions, about 87% of the talents in the artificial intelligence and data science competitions come from science and engineering backgrounds. Since economics and management are closely related to data, many of the competition questions come from the financial field, so some contestants come from economics and management majors.
occupational characteristics
Students are the main force of competition talents, accounting for about 64%. Most of the employees use artificial intelligence and data science competitions as skill training grounds and interest clubs. The main groups are workers in IT and related industries, and those engaged in finance, consulting, etc. related workers.
regional characteristics
Contestant appeal
Classification and summary based on the typical characteristics of the participants can help us better design the event and analyze and judge the development direction of the future event.
-
Bonus Reputation Claims:
Participants whose core demands are bonus incentives and reputation acquisition are mainly current students and newcomers in the workplace. They have plenty of time at their disposal, high technical strength, and the energy and strength to impact bonuses and TOP rankings. An important part of the front row players.
-
Employment and entrepreneurship demands:
Participants with employment and entrepreneurship as their core demands are highly targeted and mainly participate in professional competitions in certain vertical fields, such as finance, technology, and biomedicine.
-
Learning improvement demands:
Participants whose core demands are data acquisition and skill training are mainly beginners in data science, who choose to participate due to difficulties in data acquisition and lack of training opportunities. They are the highest proportion of participants in major competitions. These contestants are weak in technical strength, but have a strong willingness to learn and have the opportunity to develop into the core players of the competition.
-
Social activity demands:
The number of contestants with social activities as the core appeal is small and the influence is great. Under the environment of accelerated development of the competition community, some in-depth contestants have begun to form communication communities, clubs, and self-media platforms, becoming opinion leaders in the competition field, exerting a huge influence on competition operation recruitment and public opinion trends.
Dismantling of the basic structure of the competition
Outlook
1. Balance the regional differences in the construction of digital China and empower the ecological development of digital government.
2. The layout of Digital China is unfolding, and competitions related to data elements may be included in the assessment indicators.
3. Cases of achievements have emerged, and the value conversion path has gradually become clear.
4. Solve the pain points of talent recruitment, and use the scoring system in the assessment operation process to accurately evaluate skills.
5. Scientific research and teaching provide landing scenarios for micro-events.
6. AIGC brings AI revolution, and NLP competition questions will become a hot spot in the new stage.
7. Paste the simulation questions into practical problems, and strengthen learning to achieve the optimal solution or become a popular type of competition questions.
8. Online events favor the cloud environment, while offline events pursue a sense of competition.
Full HD version available
DC Artificial Intelligence and Data Science Competition Platform for Download
DataCastle - Data Science Innovation and Practice Platform www.datacastle.cn/index.html
Or pay attention to our official account
<DataCastle>
Backstage reply "White Paper Overview" to receive