Recently, ChatGPT has been popular. Before that, Wang Huiwen, the co-founder of Meituan, released the AI ? ? hero list, announced that he would pay for himself, and set up Beijing Lightyear Technology Co., Ltd. to confirm his entry. Later, there was an extreme dialogue between programmers from big factories and ChatGPT. From emotional consultation, project management to novel creation, ChatGPT was almost omnipotent and omnipotent (PS, Xiaobian was trembling with fear).
Behind the explosion of ChatGPT is everyone’s embarrassment and concern for the field of artificial intelligence. So is there any application of artificial intelligence technology in the field of pig breeding? How do they combine?
The Origin and Development of Artificial Intelligence Technology
Artificial intelligence technology is a branch of computer science. The original intention of its creation is that scientists hope that computers can imitate human intelligence, so that machines can handle complex things.
Artificial intelligence can be traced back to 1950, and Allen Matheson Turing put forward the famous "Turing Test" [which refers to asking questions to the testee through some devices (such as keyboards) when the testee is separated from the testee (a person and a machine). After many tests, if the machine makes the average participant make more than 30% misjudgment, then the machine has passed the test and is considered to have human intelligence.
In the same year, he published the paper "Computing Machine and Intelligence", and put forward and tried to answer the question "Can a machine think?". After the paper was published, it received extensive attention and discussion, and Turing was later called "the father of artificial intelligence".
In 1956, John McCarthy (computer scientist and cognitive scientist), an assistant professor in the Department of Mathematics at Dartmouth College, located in the small town of Hannos in the eastern United States, invited a group of big coffee scholars including Marvin Minsky (winner of Turing Prize in 1969) and Claude Shannon (founder of information theory) to hold an academic conference. The conference mainly discussed topics such as machine imitating human intelligence, including: how to program computers, neural networks, calculation scale theory, mechanical theory (referring to self-learning), randomness and creativity.
Dartmouth Conference has been held for more than two months. Although the participants did not reach an agreement, they agreed on a word for the discussion: Artificial Intelligence (AI). At this point, the word artificial intelligence began to appear in people’s field of vision, and 1956 was also called the first year of artificial intelligence. After that, the theoretical research and practical application in the field of artificial intelligence continued to break through (see the development history of AI for details).
(Image from Demeo Consulting Wang Wei)
Before ChatGPT, the last time artificial intelligence was widely concerned was in May 2017, when Alphago defeated Li Shishi, the world Go champion, by a score of 4: 1, and will face Ke Jie, a player from China, at the world internet conference in Wuzhen. You know, after Li’s defeat in artificial intelligence, Ke Jie made a statement in the media: Even if Alpha Dog beats Li Shishi, it can’t beat me.
Before the start of the competition, Ke Jie had high hopes and was once regarded as "the last hope of mankind". However, the reality is cruel. Alphago(Master) beat Ke Jie, a talented teenager in China, with a score of 3: 0.
After the on-site interview, Ke Jie once choked:Playing chess with AlphaGo is too painful, AlphaGo is too calm, it is too perfect, and I can’t see any hope of victory.. So why is "artificial intelligence" technology so powerful?
Analysis of artificial intelligence technology
Artificial intelligence mainly includes five core technologies, namely computer vision, machine learning, natural language processing, robotics and biometrics.
Computer vision is a science that studies how to make machines "see", which means that cameras and computers are used to identify, track and measure targets instead of human eyes, and further graphic processing is carried out, so that computers can be processed into images that are more suitable for human eyes to observe or send to instruments for detection.
Machine learning is the core of artificial intelligence technology, which enables intelligent machines to simulate human behavior independently with the support of algorithm complexity theory, convex analysis, statistics and other disciplines. Machine learning refers to how to improve the performance of specific algorithms in empirical learning, that is to say, machine learning is based on massive data or past experience to optimize the performance standards of computer programs.
To put it simply, this process is similar to personal self-reflection, which is to review past experiences and then adjust the optimization behavior so as to do better next time. But different from personal reflection, personal reflection has limited sources of experience, while machine learning is based on a huge database given by developers, with wider sources of experience and data and more timely feedback based on goals.
Natural Language Processing (NLP) is a variety of theories and methods to study how to achieve effective communication between people and computers in natural language. Natural language processing is a science integrating linguistics, computer science and mathematics, which can be mainly applied to machine translation, public opinion monitoring, automatic summarization, viewpoint extraction, text classification, question answering and so on.
Robot refers to performing tasks such as working or moving through programming and automatic control. Robots have the basic characteristics of perception, decision-making, execution, etc., which can assist or even replace human beings to complete dangerous, heavy and complicated work and improve work efficiency and quality. At present, sweeping robots have entered the daily life of the public.
Biometric technology is closely combined with high-tech means such as optics, acoustics, biosensors and biostatistics through computers, and uses the inherent physiological characteristics (such as fingerprints, faces, irises, etc.) and behavioral characteristics (such as handwriting, voice, gait, etc.) of human bodies/creatures to identify their identities, such as "face recognition" and "pig face recognition".
Therefore, artificial intelligence can be understood as imitating human information input (images, words, sounds, etc.), information processing (based on the correct thinking model in the past) and then action execution, and constantly strengthening various abilities in the process to achieve the set goals and continue to improve. The core of artificial intelligence lies in deep learning, that is, continuous feedback and continuous optimization based on strategy.
The mechanism of deep learning is similar to the deliberate practice learning method proposed by Florida psychologist Anders Millard J. Erickson. Anders pointed out that the key factor to distinguish a person’s mediocrity and Excellence in the professional field is the degree of deliberate practice. The longer the deliberate practice, the higher the professional level. Deliberate exercises mainly include four elements:Clear goals, staying away from the comfort zone, concentration and timely feedback.. Different from human deliberate practice, the deliberate practice process of the machine has no emotion, so it is more efficient to execute.
Just a few months after Alphago(Master) defeated Ke Jie, in October of 17, DeepMind team published a new paper in Nature magazine, and launched a new generation of product AlphaGo( Zero). The paper points out that Alphago(Zero) reached the level of Alphago(Master) in only 21 days, and when Alphago(Zero) played the 40th day, it had already defeated all previous programs and won the world Go championship.
Afterwards, the industry put forward three cores for Alphago to quickly reach the top level in the world: first, it adopted a learning algorithm combining machine learning and neuroscience; Second, in Google’s powerful cloud computing system, more than 30 million steps of professional chess player’s chess manual have been learned through a large amount of data analysis; Third, throughThe ever-increasing self-gameFind a better idea than the basic chess manual. Artificial intelligence technology is created by human beings, and the achievements in some aspects are far beyond human beings, which may also be worth thinking and learning.
Application of artificial intelligence technology in pig farm
The combination of artificial intelligence technology and agriculture can be traced back to March 2015. Li Keqiang, then Premier of the State Council, put forward the action plan of "internet plus" for the first time in his government work report; In July of the same year, the State Council issued the Action Plan on Actively Promoting the internet plus (hereinafter referred to as the Action Plan).
The Action Plan clearly puts forward to actively promote the modern agriculture in internet plus, and points out that it is necessary to establish standardized scale livestock and poultry breeding bases and aquatic healthy breeding demonstration bases,Promote the popularization and interconnection of intelligent devices such as accurate feed delivery, automatic disease diagnosis and automatic waste recycling.. The release of Action Plan opened the historical curtain of agriculture in internet plus.
After that, AI technology, like other emerging Internet technologies, was gradually applied to the practice of agricultural industry. On the one hand, new cutting-edge technology, on the other hand, traditional agriculture, how do they combine? What kind of sparks will break out?
Scenario 1: Pig farm monitoring and pig inventory
Since the outbreak of African swine fever in 2018, biosafety has been crucial for aquaculture enterprises. Under the traditional breeding mode, people and things inside and outside the pig farm are complicated, and materials, vehicles, birds and animals come in and out frequently. It is extremely difficult and time-consuming to achieve effective supervision. For example, in the process of dissecting dead pigs, improper wearing of protective clothing by operators may cause the spread of diseases in pig farms. It is difficult to pass the manual on-site review, and it is impossible to supervise all the time. By using technologies such as artificial intelligence and Internet of Things, employees’ misconduct can be identified and early warning signals can be conveyed to managers in time, which can effectively avoid the risk of disease spread.
In addition, artificial intelligence technology can also be used to count and estimate the number of pigs. In the traditional breeding process, pig dealers need to go to the pigsty for on-site counting and weighing in the pig selling process. By using the camera above the pig farm combined with artificial intelligence technology, the pigs in the pen can be counted in real time, so as to achieve the effect of remote pig watching and real-time weight estimation.
The AI patrol and early warning interface of "Pig Xiao Zhi" of rural credit cooperatives.
Scheme of intelligent pig farm with integrated rural credit, mathematics and intelligence for message or private message consultation
Scene 2: Intelligent dung cleaning robot and inspection robot
Pig farm is a model of large-scale farming. Large-scale farming and high manure output are difficult problems that plague the operation of enterprises. In traditional farming mode, manual cleaning is usually needed, and the labor input cost is high in the process of manual cleaning. Some pig farms use water to clean manure, which consumes a lot of water, and at the same time, it is easy to cause high humidity in pig houses and cause health problems of pigs. By using artificial intelligence technology to develop a dung sweeping robot, it can effectively solve the problems of difficult cleaning, reducing water consumption and saving costs.
Similarly, robots developed by combining intelligent speech recognition and visual recognition technologies can cruise around the pigsty all day, find abnormal pigs (such as fever and shivering) in the pigsty in time, prevent and control diseases in advance, and effectively resolve the risk of disease infection.
Mu Yuan inspection robot
The above are only typical application scenarios of artificial intelligence technology in pig farms. With the expansion of business scale and the improvement of cost control and biosafety requirements of enterprises, the combination of artificial intelligence technology and pig breeding is becoming closer and closer.
At last year’s Deep Bay Meeting, in view of the application of intelligence in pig farms, Qin Yinglin, chairman of Mu Yuan Co., pointed out: "It takes three years for us to train an excellent employee, and many employees are unqualified for three years, which takes five or even ten years. But we want to make a machine in the assembly line, and gather all our wisdom of raising pigs. We are producing this very quickly now, and producing thousands of units a day is equivalent to producing the corresponding number of laborers. "
It can be seen that artificial intelligence technology has been applied on a large scale in the farms of some enterprises. With the continuous development and mature application of technology, it is certain that "unmanned pig farms" and "unmanned farms" will eventually come true.
Talking about the development of AI from ChatGPT, consulting with Deme.
Nature is heavy: human beings have misunderstood the dopamine mechanism in the brain! ? Top Edition AlphaGo Enlightens Brain Science, Quantum Bit
Qin Yinglin’s Sharing in Deep Bay Meeting, Wandou Agricultural Science.
Guiding Opinions on Actively Promoting "internet plus" Action, the State Council