Tal won the champion of neurips2020 international competition

Time:2021-10-6

Recently, the international artificial intelligence top-level conference neurips 2020 (2020 Conference on neural information processing systems, that is, the conference on neural information processing systems, hereinafter referred to as “neurips 2020”) announced the competition results.

Tal AI middle platform machine learning team won a single champion, a single runner up and a comprehensive evaluation runner up in the four single tasks of neurips 2020 “education challenge”, showing the world the frontier exploration of China’s “Ai + education”.

As one of the world’s top conferences in the field of artificial intelligence and machine learning, neurips aims to promote the research and exchange of neural information processing systems in biology, technology, mathematics and theory, which has attracted much attention inside and outside the industry. This year’s neurips held an education related challenge for the first time, which attracted more than 80 professional teams from Peking University, China University of science and technology, Pennsylvania State University and so on.
Tal won the champion of neurips2020 international competition

(neurips 2020 is one of the top international artificial intelligence conferences)

Won the international top competition champion and demonstrated the strength of Tal’s “Ai + education”

Neurips2020 “educational challenge” competition closely focuses on the “diagnostic problems” in education, and requires the participating teams to analyze the response records of students’ nearly 1000 mathematical multiple-choice questions in the past two years, so as to accurately predict students’ answers and which questions have high response quality, and determine a personalized question sequence that can best predict their answers for each student.

For the competition questions, the machine learning team in tal AI adopts the most cutting-edge pre training language model in natural language processing (NLP) to transform the competition questions into adaptive learning problems for prediction, and effectively uses the sequence information of students’ practice. Finally, in the task of “evaluating the quality of questions” In the task of “ability diagnosis of new students through a very small number of questions” and comprehensive ranking, they won the champion, runner up and runner up respectively.

Tal won the champion of neurips2020 international competition

(award winning Email of Tal AI middle machine learning team)

With the deep application of digital technology in education, the global students’ demand for personalized and high-quality educational resources is becoming more and more urgent. As the International Ai summit conference, the research on the results of neurips 2020 “education challenge” competition in the field of AI is likely to have a sustained and real impact on the personalized education of millions of students around the world. Through long-term research and accumulation in the field of NLP and data mining, tal AI middle machine learning team can more accurately understand students’ learning status and give personalized learning suggestions, which lays a good foundation for the application of data analysis, content recognition, automatic correction and other technologies in educational scenes.

Increase investment in scientific research and talents, and promote the implementation and application of “Ai + education”

With solid technical accumulation and in-depth understanding of smart education, tal was approved by the Ministry of science and technology in August 2019 to undertake the construction of a new generation of national artificial intelligence open innovation platform for smart education, promote the openness and sharing of the industry and help the growth of small, medium-sized and micro enterprises driven by the major needs of smart education. At the same time, tal has continuously strengthened the construction of bottom scientific research capacity, established joint laboratories or carried out joint scientific research projects with 6 universities and institutes such as Tsinghua University and the Institute of computing of the Chinese Academy of Sciences, and continuously strengthened scientific research and research and academic cooperation.

At the same time, tal has continuously increased the training of educational talents, bringing together a large number of high-tech talents with educational ideals and committed to the pursuit of love and science and technology to make education better. Dozens of academic achievements participated by TAL AI team were selected into international top academic conferences such as neurips, AAAI, WWW, emnlp, aied and NCME, and successively won the champion of CVPR emotionet competition, Ubicomp competition, ccl2020 competition and other honors. At the AAAI summit conference in New York, tal AI team successfully organized the first AI for Education Symposium to promote international academic exchanges in the field of “Ai + education”.

While increasing investment in scientific research and talents, tal is also constantly applying AI technology to real education scenes. In recent years, tal has continuously made breakthroughs in cutting-edge core technologies, realized the product application of a number of technologies based on the continuous accumulation of AI capabilities such as voice technology, visual understanding and knowledge map, and created innovative product solutions including AI classroom, teaching process evaluation, oral expression ability evaluation, homework shooting, search and correction. At present, tal has explored and formed more than 100 AI abilities and more than 10 AI solutions for educational scenarios covering all teaching links of “teaching, learning, testing, practice and evaluation”.

For example, the “Chinese and English oral expression ability evaluation solution” independently developed by TAL, based on the industry-leading infant speech recognition technology and multi-dimensional oral expression evaluation algorithm, outputs evaluation reports from the dimensions of fluency, emotion, content relevance, semantic logic and language use, which is not only widely used in tal’s brands such as XRS Peiyou and XRS online school, It is also open to other educational practitioners in the industry. Tal’s photo search software takes “questions can’t be found, teachers answer for free” as the core function. It has powerful AI technology and a large number of high-quality question banks. One click photography can quickly obtain problem-solving ideas and detailed step analysis, and continue to launch more intelligent AI functions.

It has successively won international authoritative competition awards, which marks the recognition of Tal’s scientific research strength by the international academic community, and also highlights the practical value of Tal AI technology in the educational scene. In the future, tal will continue to promote in-depth research in the field of educational science and technology, and provide educational practitioners with full scene, full process and full cycle technical and service support based on the national new generation artificial intelligence open innovation platform of intelligent education, so as to provide children with a fairer and more quality educational experience.

Recommended Today

Solve a problem with responsewriter in golang

When using the three methods set / writeheader / write in context.responsewriter, the order of use must be as follows, otherwise a setting will not take effect. ctx.ResponseWriter.Header().Set(“Content-type”, “application/text”) ctx.ResponseWriter.WriteHeader(403) ctx.ResponseWriter.Write([]byte(resp)) As shown in Figure 1: ctx.ResponseWriter.Header().Set(“Content-type”, “application/text”) ctx.ResponseWriter.Write([]byte(resp)) ctx.ResponseWriter.WriteHeader(403) It will cause the return code to be 200 all the time Supplement: in go, […]