In recent years, digital technologies represented by AI (artificial intelligence) and cloud computing have gradually become an important engine driving the high-quality development of the real industry and the digital economy.
Technological innovation achievements have become the best way to help solve social pain points and deliver warmth through the form of technological inclusiveness.
A few days ago, the 2nd Tencent Light Public Welfare Innovation Challenge co-hosted by Tencent and China Children's Center held a final roadshow.
During the competition, many young teams showed their charity ideas of "warming the world".
Build a "canteen" for beavers
The cute-looking beaver is an amphibious mammal.
Beavers living in China are called "Mengxin Beavers", distributed in the Altay region of Xinjiang Uygur Autonomous Region, and are national first-class protected animals.
In 2018, the population of Mengxin beaver in China was only about 500, and it was at risk of extinction.
It was also in that year that Chu Wenwen, a "post-90s" girl who graduated from Beijing Forestry University, came to Altay and took over the task of protecting Mengxin beavers from her father.
In order to protect the beavers, Chu Wenwen and the volunteers established the Altay Regional Nature Conservation Association to establish a systematic information database for the Mongolian beaver and design a systematic protection plan.
The survival of Mengxin beavers is inseparable from shrubs. Shrubs are not only the main food source of beavers, but also the "sanctuaries" for beavers and other wild animals to reproduce, live and avoid natural enemies.
Since shrub plants are not commercial crops, they were previously planted in very small quantities in Altay.
To this end, Chu Wenwen and members of the association launched the "Beaver Canteen" project to restore the shrub vegetation community in a large area in the Ulungu River Basin in Altay, and planted more than 600,000 shrubs and willow saplings.
Growing shrub willows is by no means a one-shot deal.
Where the seed is and the effect of the seed is inseparable from scientific research and judgment.
The team members introduced that before and after planting trees, it is necessary to observe and analyze changes in the regional habitat and biodiversity in the area - what animals have appeared in this area?
Is the frequency high or low?
Are species added or lost?
Wang Dapeng, a member of the Nature Conservation Association in Altay, said that in order to observe the biodiversity changes of habitats for a long time, the team has set up more than 500 infrared cameras in the field for observation, and the video capacity of observation in one year reaches five or six hundred terabytes (T) .
In the beginning, the team had to use the "clumsy method" of human identification to record the frequency of the different animals in the video.
More than a dozen graduate students and more than 200 online volunteers often take two to three months or even half a year to complete the voluminous data analysis, and the accuracy rate is difficult to guarantee.
Under the recommendation of the scientific research team of Beijing Forestry University, the team members used AI technologies such as Tencent Cloud Image Recognition to quickly identify the animal species captured by the camera.
"Just like face recognition, AI also has functions such as 'bird face recognition' and 'leopard face recognition' after 'hard' learning and understanding. Our work efficiency is immediately improved." Wang Dapeng said, in With the help of AI technology, the manual review from several months was shortened to two days, which greatly improved the protection efficiency of beavers.
With the efforts of the team, the population of the new beaver in Altaymont has increased from 500 to about 620 in just a few years, an increase of 20%.
Not only that, but based on the restoration of wildlife habitats, the number of local tiger ferrets, great egrets and other wild animals is also growing rapidly.
"In the field of biodiversity protection, the technical support of AI has allowed us to achieve breakthroughs. Now, we can proudly present the results to the United Nations." Wang Dapeng said.
At the end of 2021, Chu Wenwen attended the opening ceremony of the fifteenth meeting of the Conference of the Parties to the United Nations Convention on Biological Diversity (COP15), telling the world the story of Chinese youth protecting beavers.
Take a picture to measure jaundice
In 2021, there will be about 10.62 million newborn babies in China, of which about 5.13 million have jaundice symptoms.
Although the symptoms of jaundice in the vast majority of newborns will gradually subside, severe jaundice can cause sequelae such as movement disorders, mental retardation, language disorders, and epilepsy. The monitoring of jaundice is very important.
Previously, Song Xia from Yuezhi AI Lab did not know much about jaundice.
It wasn't until the teammates' children were plagued by jaundice that a significant number of newborns were found to suffer from jaundice.
Can you take advantage of technology and do something for parents of new babies?
The team members found that the current determination of the baby's jaundice degree is usually detected by blood test or jaundice meter. Both methods need to carry the child to the hospital, which can easily cause congestion of medical resources.
Babies with jaundice will appear "yellow staining" due to the deposition of bilirubin in the skin.
Can we use AI image recognition technology to detect jaundice and relieve parents from worries?
Song Xia and the team members decided to give it a try.
Two years ago, team members developed an online jaundice test applet/App "Health Bear". After parents take "four consecutive shots" of relevant parts of the baby's body and upload pictures, AI can determine whether the newborn has jaundice disease. and intelligent judgment of its severity.
Song Xia said that "Health Bear"'s judgment of jaundice not only relies on image recognition, but also adds the precipitation knowledge base of common jaundice symptoms provided by medical institutions to the algorithm model, which improves the accuracy of jaundice recognition.
In order to collect more samples and improve the AI's learning and recognition capabilities, Song Xia and his colleagues ran around for cooperation, and used the time outside the main business to "code" and optimize the algorithm.
As the number of samples increases, the AI recognition accuracy continues to improve.
Song Xia said that the current sample size of the product has reached 100,000. After connecting to the Tencent Cloud interface and optimizing the detection, the recognition accuracy rate has increased to about 92% to 97%.
At present, the product has been connected to the official account of the Guangzhou Municipal Health and Health Commission, "Guangzhou Health Link", as a remote tool for the follow-up of children with jaundice. The identification results will be synchronized to the partner hospital, and the doctor will make the final treatment decision.
"Today, the frequency of use of the product is close to 10 million times. We hope that more parents can enjoy more accurate jaundice identification, and provide better solutions for babies through the technical means of the hospital." Song Xia said.
Drive the development of public welfare undertakings
It is understood that this public welfare innovation challenge has set three contest topics: biodiversity protection/education, minor safety protection/education and literacy education.
In the final roadshow, a total of 24 teams demonstrated public welfare projects developed using products such as Tencent Cloud AI——
The "Bean Flour Dough" project team has built an "intelligent community stray cat live broadcast service station" in urban communities.
On the one hand, stray cats can be "settled" in the service station, so that the surrounding residents can help them in a timely manner; on the other hand, popularize scientific pet raising knowledge through online platforms to help stray cats get adopted and return to their families;
The "Pupil Protection Program" team has developed an "intelligent eye care defense system" in response to the increasing myopia rate among Chinese teenagers. Based on software and hardware such as small programs, APPs, and children's watches, electronic vision charts are used to test children's vision in real time and accurately.
At the same time, the switch can be remotely controlled according to the length of time the child uses the mobile phone;
The "North Star Star" team has long been concerned about the education of left-behind children in Mandarin. The college students in the team used cloud speech recognition technology to design the "Language Ocean Bubble" applet, so that children left-behind in the countryside can learn Mandarin efficiently through follow-up reading and other methods. .
At present, the product has been piloted in Litang, Sichuan, and also supports the exchange of Tibetan-Chinese/Tibetan-English content.
Wu Yunsheng, vice president of Tencent Cloud and general manager of Tencent Youtu Lab, said that compared with the first competition last year, this year's participating teams have made significant progress in the breadth, depth and completeness of their creativity.
The greatest value of AI is "serving people". As AI technology is implemented in various industries, it will also be fed back to public welfare platforms.
Liu Yao