月度归档 2024年3月31日

　　Photo courtesy of vision china

　　Ai growing pains

　　Editor’s note In recent years, artificial intelligence has developed rapidly and its application fields have been expanding. However, at the same time, AI "rollover" cases are frequently searched. Based on this, this edition opens the column "AI Growing Pains" to focus on AI "rollover" events, analyze the phenomenon, analyze the reasons, explore the solutions, and look forward to its growth.

　　Intern reporter Dai Xiaopei

　　After the game, it is probably unprecedented for fans to ask referees to wear hats or wigs.

　　At the end of October, Scottish fans experienced an "unforgettable" football match. In the Scottish Football Champions League match between Furness and Elven, the AI cameras on the sidelines turned a blind eye to whether the players passed the ball or attacked with the ball. Instead, they followed a linesman and took a close-up of "C position" from time to time. It turned out that the AI camera mistakenly identified the referee’s bald head as a football, so he chased the whole game crazily.

　　Fans who watched the 90-minute game at home spent most of their time not watching the ball, but watching the bald heads. Many netizens joked that this game really "saw a loneliness".

　　Why does the AI camera regard the bald head as a football? What do we need to do to prevent AI from making similar mistakes? Can you think that AI is "weak" when there is a "rollover" incident?

　　Unconscious "provocation" from bald referees

　　From direct participation in sports events to recording athletes’ performance, to live broadcast of competitions and analysis of athletes’ health status, AI is becoming the darling of the sports world. A few months ago, Barcelona Football Club (Barcelona) also joined hands with video technology company Pixellot to create an artificial intelligence coaching solution.

　　Unexpectedly, AI, which has made great strides in sports, unexpectedly encountered an unconscious "provocation" from a bald linesman. Because the linesman’s bald head is too bright and the sun shines, the AI camera can’t tell which is the ball and which is the head. Previously, because the Furness team said that the AI tracking technology they used could clearly transmit the live images to the homes of every season ticket buyer, so that fans who could not go to the home due to the COVID-19 epidemic would not miss any game.

　　It is reported that the camera used in the live broadcast of this competition is the multi-camera system provided by Pixellot, which cooperates with Barcelona. The system is powered by NVIDIA’s graphics processor (NVIDIA GPU), and the captured video resolution can reach 8K. These cameras can be installed in a fixed position without being manipulated by the camera operator. In order to capture the key moments, Pixellot collected hundreds of thousands of hours of sports videos to train its algorithm on the NVIDIA GPU of the local workstation.

　　With a huge amount of available data, using deep learning algorithms and high-performance GPU computing blessing, it has three major driving forces to push AI forward. Why did Pixellot’s AI camera "roll over"?

　　After the event, the relevant clubs and the technology companies that make cameras reflected. The problem seems very clear: the size and shape of football are similar to those of a human head, and the direct sunlight makes the AI camera fall into "confusion". I hope this will not happen again, because the Furness team responded that it knew the problem and will improve the next game to bring a better experience to the audience.

　　Pixellot also said that it is not difficult to solve this problem. The existing target detection and tracking technology is mature. Pixellot didn’t consider the influence of bald head in the design stage, so it needs to collect some data of football and bald head to fine-tune the algorithm to eliminate the interference from bald head.

　　Some technicians said that not only "this is a ball" data set, but also a "this is not a ball" data set is needed when training the AI camera of the live ball game. Bald head, bright white shoes, lighting, balls on the training ground next to the competition field and balls used by players to warm up are all interference factors that need to be considered when training AI.

　　AI "poor eyesight" is the norm.

　　Although the performance of AI cameras can be improved by increasing data "feeding", strengthening training and improving algorithms, some professionals believe that with the gradual expansion of AI application scenarios, such "rollover" incidents will exist for a long time.

　　“AI‘ Rollover ’ It’s normal, no ‘ Rollover ’ It’s strange. " Professor eecs and Huang Tiejun, president of Beijing Zhiyuan Artificial Intelligence Research Institute, spoke frankly when interviewed by Science and Technology Daily reporter.

　　Huang Tiejun believes that on the surface, the failure of the AI ? ? camera may be due to insufficient pre-training, but the main reason is that the current computer identification system is only trained with specific data. For example, in the above example, the neural network trained with a large number of football videos has surpassed humans in identifying football, but it ignores that this network is more sensitive to bald heads.

　　It is common to recognize or "turn a blind eye" unknown objects at random.

　　Machine vision is to endow the machine with visual perception, so that the machine has the scene perception ability similar to that of biological vision system, which involves optical imaging, image processing, analysis and recognition, execution and other components.

　　"Take the camera as AI ‘ Eyes ’ In real scenes, there is still a long way to go for AI to recognize football and baldness like human eyes. " Huang Tiejun said.

　　When can I finish this road and even realize that the eye of AI surpasses the human eye?

　　It depends on when machine vision will bridge the gap with biological vision. "The brain in the skull senses the outside world in real time through more than 3 million nerve fibers, of which there are more than 1 million behind each eye." Huang Tiejun said, "The machine vision developed to this day is still a drop in the bucket compared with the biological vision system that took hundreds of millions of years to evolve."

　　The human eye has strong adaptability, can identify targets in complex and changing environments, has advanced intelligence, and can use logical analysis and reasoning ability to identify changing targets and summarize laws. On the other hand, machine vision can use artificial intelligence neural network technology, but it can’t identify the changing target well. Due to the constraints of hardware conditions, the current general image acquisition system has poor color resolution.

　　"Compared with biological visual neural networks, the visual neural networks of artificial intelligence are far apart in structure and scale, so their functions are much worse." Huang Tiejun said, "In real applications, machine vision ‘ Rollover ’ It is not an accident. It is only a case to identify the bald head as a football. Similar problems actually exist in large numbers. "

　　Huang Tiejun said: "This time, the technology provider can make up for the loophole that the bald head mistakenly thinks of football, but there are still more loopholes. Deceiving the face recognition system with antagonistic picture training is only the tip of the iceberg."

　　Different technical routes are racing.

　　"Machine vision based on deep learning has made great progress in image recognition, but it has not really solved the perception problem." Huang Tiejun believes that deep learning is far from grasping the complexity of human visual system.

　　Deep learning is based on the training of image and video big data, which is far from the biological vision that actively perceives the dynamic world, and it is still not out of demand for computing power. For example, if the video frame rate is increased from 30 to 30,000, the computing power of deep learning needs to be increased by 1,000 times.

　　Biological neural network is a pulse neural network, which is more suitable for completing visual information processing. Huang Tiejun believes that learning from the neural network structure and information processing mechanism of biological vision system and establishing a new set of brain-like vision information processing theory and technology is the hope of restarting machine vision.

　　Experts said that there are two main technical routes to develop artificial machine vision. One is to build a powerful intelligent system by collecting more data, increasing the amount of data and strengthening training; The second is to imitate the biological nervous system, follow the gourd painting gourd ladle, clarify the structure and even mechanism of the biological nervous system, and develop future intelligence on this basis.

　　Huang Tiejun believes that the second path is more effective than the first one. "In a short time, the first one is easier to achieve results. But in the long run, it is more direct to start with biological neural networks and more confident to achieve the goal. "

　　At present, most AI scholars support the first path, that is, to develop artificial intelligence including machine vision through "big data+big computing power". Huang Tiejun set foot on the road taken by a few people because he firmly believed that the biological visual neural network had great potential to be tapped. "Biological brain is the product of hundreds of millions of years of evolution, and it is the best transcendental structure. Powerful intelligence must rely on complex structures and stand on the shoulders of evolution. It seems difficult, but it is the fastest. "

　　Turing, the "father of computer science", has long expressed his admiration for the biological brain. At the beginning of 1943, Shannon suggested that "cultural things" could be instilled in the electronic brain. Turing once retorted in public: "No, I’m not interested in building a powerful brain. What I want is just an ordinary brain, just like the head of the chairman of AT&T."

　　Is developing machine vision, or artificial intelligence, starting from scratch, imitating biological neural networks, or is there another way? There is no conclusion yet. On different tracks, artificial intelligence is accelerating.

　　Although there are many cases of AI "rollover", such as the practicability of Google AI eye disease diagnosis system is greatly reduced in Thailand, Tencent AI translation made a joke at the Boao Forum in 2018. But the AI era is accelerating and unstoppable.

　　"It must be admitted that AI has indeed solved many practical problems, and it will gradually replace some functions of people. But we can’t exaggerate it too much. It is still far from our imaginary intelligence and needs more breakthroughs. " Huang Tiejun said that to keep an open mind and bridge the gap between "experimental simulation" and "real world", the development of AI technology has a long way to go.

　　Baked starch sausage

　　It is a "online celebrity snack" that many people love to eat.

　　But recently,

　　The starch sausage was "collapsed"

　　Baked starch sausage at roadside stalls all over the country

　　The price is generally 2-3 yuan/root.

　　In the wholesale market

　　The price of starch sausage is about 0.5-0.6 yuan per root.

　　The cost price of factory production is lower.

　　At this price.

　　Is there meat in the starch sausage?

　　What kind of meat, if any?

　　If there is no meat

　　Where does the meat flavor of starch sausage come from?

　　Five common starch sausage products bought by reporters in the market

　　The meat in the starch sausage is mostly bone mud.

　　Commonly used in pet food production.

　　A comprehensive observation of the five brands of starch sausage with high sales volume on the market at present shows that there is little difference in the ingredient list, and they are basically composed of meat, water, starch, white sugar, monosodium glutamate, spices and other seasonings, as well as additives such as sodium tripolyphosphate, sodium pyrophosphate, sodium hexametaphosphate, sodium carboxymethyl cellulose and temptation red, totaling 15 or 6 components.

　　List of ingredients for starch sausage products

　　According to the General Rules for Labeling in prepackaged foods, all ingredients should be arranged one by one in descending order of adding amount when manufacturing or processing food, but the "starch sausage" ranked first in the ingredient list is meat.

　　The staff of a starch sausage manufacturer in Henan revealed that they all use chicken and chicken bone paste, and duck meat is cheaper, and most of them are starch.

　　At present, most of the people who sell chicken bone paste are merchants who operate pet food, and the merchants said: "Bone paste is not recommended for people to eat, and barbecue is not good."

　　The picture shows the "chicken bone paste" product searched and displayed on the e-commerce website.

　　He Jiguo, an associate professor in the College of Food Science and Nutritional Engineering of China Agricultural University, once said that the use of chicken neck mud and chicken rack mud instead of the original meat is essentially adulterated and shoddy, and the nutrients in it can’t be absorbed by the human body.

　　The staff of the starch sausage manufacturer also revealed that the ex-factory price of starch sausage is 80 yuan/box, and each box contains 120 pieces, totaling 24 kg. If you get more goods, you will get a gift, which is basically 50 cents to 60 cents.

　　Starch sausage has less meat but full of "meaty taste"

　　There is essence blessing.

　　It is difficult for consumers to tell what kind of meat is in the starch sausage when they eat it.

　　On the one hand, because the amount of starch sausage meat is small; On the other hand, an important ingredient in the seasoning of starch sausage is essence. Judging from the samples of the five brands of starch sausage purchased, all the ingredients have flavors. The staff of the above-mentioned Henan starch sausage enterprise also directly said, "Don’t worry about less meat and no meat flavor, there are flavors.

　　Henan Shuangrun Food Co., Ltd. Flavor Barbecue Pork Sausage Enterprise Standard Screenshot

　　There is no special national standard for starch sausage

　　Although it looks like ham sausage, starch sausage is not produced according to the national standard of ham sausage, and there is a certain gap between the physical and chemical indexes of the product and ham sausage, so it can not be regarded as ham sausage product.

　　According to the national standard, the sausage is in accordance with the General Rules for the Quality of Ham Sausage (GB/T 20712-2022). In terms of physical and chemical requirements, the starch contents of ordinary, excellent and super ham sausages are ≤10.0g/100g, ≤8.0g/100g≤6.0g/100g respectively.

　　Taking the enterprise standard "Sausage" (Q/PFY 0006S-2023) of Panshan Fuyu Food Co., Ltd. as an example, the starch contents of three products, namely sausage, skin sausage and vermicelli sausage, are ≤10%≤10% and ≤25% respectively.

　　However, other starch sausage manufacturers did not even limit the starch content of starch sausages.

　　So, is there enough meat and corresponding protein in the starch sausage?

　　As it turns out, there is no clear standard requirement in this respect …

　　The protein content of the sample starch sausage purchased by the reporter is 5.295.5g/1009, which is lower than the most common standard of ham sausage. However, most enterprise standards for starch sausage do not make clear requirements for protein and other indicators.

　　In addition, there is no special national standard for starch sausage, and the reference standard seems to be vague.

　　A person from a standardized technical consulting company said that to formulate enterprise standards, it is necessary to refer to the raw material list and technological process of products and then synthesize multiple national standards or industry standards. If the national standard of cooked meat products is used as the basis for enterprise standards, as long as the meat in their products meets the national standard of cooked meat products. Of course, because there is more starch in starch sausage, you can also choose to use the national standard of starch products as the basis of enterprise standard.

　　According to the relevant regulations, the description of enterprise standards should specify the contents that are stricter than the national or local food safety standards. However, in the enterprise standards of Panjin Yufei Food Co., Ltd., Henan Shuangrun Food Co., Ltd. and Panshan Fuyu Food Co., Ltd., only "the lead limit index is stricter than the national standard" is mentioned.

　　The standard of starch sausage has not become higher for more than ten years.

　　Water and fat in enterprise standards

　　The content is getting higher and higher

　　Although it has existed in the market for more than ten years, the standard of starch sausage has not improved obviously with the development of economy and society, but there are some signs of relaxation.

　　Screenshot of vermicelli sausage (Q/PFY 0004S-2013)

　　Screenshot of vermicelli sausage (Q/PFY 0004S-2013)

　　Comparing the two standards, we can see that there is more water and more fat under the new enterprise standard of starch sausage of Panshan Fuyu Food Co., Ltd.

　　In this regard, an expert who did not want to be named suggested that when consumers choose meat products, it is recommended to check the nutrition label and try to choose products with less fat and lower sodium content.

　　Netizens are hotly discussing:

　　The pure starch in the picture has meat?

　　A few years ago,

　　Ham sausage pursues more meat

　　Adding starch will be rejected.

　　Today, some consumers are

　　Turn your head and pursue starch sausage

　　Some people think they eat starch sausage.

　　It’s not for meat.

　　"Starch sausage, since the name is starch.

　　There is no demanding meat quantity. "

　　"The price of starch sausage is only scary when there is meat."

　　"But I don’t like eating any meat."

　　Comprehensive Guangzhou Daily, Yangguang.com, Hubei Jingshi, comments from netizens, etc.

Source: Xinmin Evening News

　　BEIJING, Beijing, March 19 (Reporter Wang Muqing) On the occasion of the second anniversary of China’s football reform, the China Football Association organized an in-depth discussion among people in the football field. Facing the reporter, Cai Zhenhua, president of the Football Association, bluntly said that after the first step of the Overall Plan is basically completed, it will move towards the second step in the future. He said that when China’s football reform enters the deep water area, it will be confident that it can embark on a football road with China characteristics.

　　The first step of foot reform has basically completed the goal of moving towards the second step in the future

　　On March 8, 2015, the General Office of the State Council officially issued the Overall Plan for Football Reform and Development in China. In the past two years, the development of football in China has gone beyond the level of sports, including the Ministry of Finance, the Ministry of Education, the National Development and Reform Commission and other ministries and commissions, which have been deeply involved in the reform of football in China.

　　Li Yuyi, executive deputy director of the State Council Inter-Ministerial Joint Conference Office and vice chairman of China Football Association, concluded: "‘ Three steps ’ The first step of the strategy has achieved initial results. " The so-called first step is clearly defined in the Master Plan: improving the environment and atmosphere of football development, rationalizing the football management system, formulating the medium-and long-term development plan of football, innovating the football management mode with China characteristics, and forming a pattern of coordinated development of football career and football industry.

　　Speaking of this achievement, Cai Zhenhua said: "The first step has been basically completed, and (in the future) it is necessary to move forward to the second step." According to the Plan, the second step includes a substantial increase in the youth football population, the organization and competition level of professional leagues reaching the first class in Asia, the national men’s soccer team being in the forefront of Asia, and the women’s soccer team returning to the ranks of world-class strong teams.

　　"No matter the colleagues in the football field, the society and the media, everyone has given great affirmation to the football reform in the past two years. However, the new contradictions of new problems arising in the development process need to be solved one by one. " Cai Zhenhua said, "We are confident that as long as we are on the right track and stick to the football reform, we believe that the hope of football in China will surely emerge. We will work hard to speed up this pace. "

　　Football reform has entered the deep water area and is confident of stepping out of the road of football with China characteristics

　　Cai Zhenhua said frankly that the reform of football in China is not a certain field, but all-round. At present, 94% of the primary tasks have achieved phased results, which can be copied and popularized. However, when the reform enters the deep water area, the difficulty will surely increase.

　　He bluntly pointed out: "From the China Football Association to the local football association to the sports department, this reform is to change your own life, so it is very difficult."

　　However, Cai Zhenhua believes that the initial results of the reform will give you confidence in the determination to deepen the reform: "As long as we respect the laws of football development, social development and market development, and learn from the experience of world football powers, we will certainly be able to combine China’s national conditions and embark on a football road with China characteristics."

　　In the face of people in the football world, Cai Zhenhua mentioned a lot of exciting information, such as the goal of adding 60,000 football venues in 2020; For example, according to the international football industry accounting for 40% of the sports industry, by 2025, the football industry in China should reach at least 2 trillion yuan. He bluntly said that how to achieve this goal will be an important symbol to measure the success of the reform.

China Securities Network News (Reporter Wan Yu) Liu Xiaolan, vice president of TCL Huaxing, a subsidiary of TCL Technology, said in an interview with china securities journal recently that on the whole, the panel industry will continue to recover in demand this year, and TCL Huaxing will need to fix production, maintain healthy inventory and optimize the supply chain system to ensure delivery when the market changes. At the same time, TCL Huaxing will continue to lay out and maintain a high growth rate in the fields of fast-growing folding screens and car screens. 　 　

Since the fourth quarter of last year, panel prices have continued to rise, and the panel industry has attracted the attention of investors. According to the research report of Huaxi Securities, the price of TV panels in early March is close to the high price in September last year. Due to the better downstream demand expectation and the early stocking of the whole machine factory, the current supply and demand are still tight. Huaxi Securities expects that the price of TV panels will still rise in the second quarter, and the price of TV panels is expected to break through the high level last year. 　 　

Regarding the trend of the panel industry, Liu Xiaolan thinks that with the moderate economic recovery in China and the downward inflation in Europe and America, the overall purchasing power of consumers is rising, and the sales of TV, notebook computers, monitors and other products will be driven by the superposition of domestic "618" and summer holidays. On the demand side of the panel, she predicts that this year will be more optimistic than last year. On the supply side, panel enterprises have generally changed their business philosophy, from pursuing full production and full sales to fixing production according to demand. As a whole, the inventory of the panel industry is relatively healthy, and the panel price is relatively stable, so there is no more ups and downs. 　

She introduced that TCL Huaxing will pay close attention to the changes in demand and ensure the flexibility of factory production, so that it can timely respond to the adjustment of production brought about by changes in the market. As a head enterprise, TCL Huaxing has built its own moat in the upstream supply chain. The company will also give full play to this advantage, continue to strengthen cooperation with the upstream supply chain, especially the strategic partners of key materials, establish strong links, share information, optimize the supply chain system, and ensure that TCL Huaxing’s factory can respond to uncertain situations in time and ensure the delivery of the whole supply chain. 　 　

TCL Huaxing is also actively deploying in the fast-growing sub-categories such as folding screens and car screens. Liu Xiaolan said that folding screen mobile phones have become a fast-growing segment of mobile phone products in recent years. TCL Huaxing has launched a number of folding screen products, and has industry-leading technologies such as LTPO and PLP. On the one hand, the company will continue to make breakthroughs in folding screen technology in the future, reducing the weight and thickness of the screen, extending the battery life and service life. On the other hand, it will start with materials and design, reduce the cost of folding screens, and make folding screens more widely used in mobile phones, tablets, laptops and other products. 　 　

TCL Huaxing’s shipments of in-vehicle screens also maintained a high growth trend. TCL Huaxing did not enter the in-vehicle field for a long time. Last year, the shipments were about 5 million pieces. Although the shipments were not large, the growth rate was large, which doubled and reached the first place in the industry. "In the past, we only made efforts on high-end car screens. At that time, we didn’t have a production line of amorphous silicon," Liu Xiaolan said. "After the mass production of t9 production line last year, we developed a lot of amorphous silicon car displays, and we will catch up with them at a very fast speed."