With the explosion of ChatGPT, artificial intelligence has become one of the most focused attention recently. Entrepreneurs are rushing to enter the market, and large factories are also recruiting, and it is not surprising that AI talents are offered an annual salary of millions. At the same time, artificial intelligence fever is also spreading within universities. In 2018, 35 domestic universities took the lead in obtaining artificial intelligence professional construction qualifications, most of which were 985 and 211 colleges and universities with strong strength, and in the following years, ordinary universities around the world have also followed suit, even including junior colleges.
To some extent, the establishment of artificial intelligence majors in colleges and universities is forward-looking. But as a new, unknown and profound major, ordinary colleges need to explore, and teachers need to learn. Young people come with the unknown and move forward with the unknown.
Now, AI can write copywriting, PPT, code, and draw. And where these young people who encounter the "wind" will go in the future, artificial intelligence majors cannot answer for the time being.
Wen | Wu Yuchen
Editor|Chu Ming
Half fiery, half confused
After lights out, it's time to tell the truth.
Zhang Tian's bedroom habitually chats about her recent troubles. Once, everyone talked about not regretting choosing the "artificial intelligence major", and four people in the room said without hesitation: "Regret."
At that time, Zhang Tian was studying at an engineering college in Xi'an until the second semester of her sophomore year, with no project, no internship, and insufficient ability, and felt a little flustered. The course is also becoming more and more difficult, and she has begun to understand the data structure she has recently learned - a few numbers to compare, from small to large, the order that humans can see at a glance, to teach the computer. But the computer is "very stupid", to teach it two numbers two numbers to look at two numbers, the two numbers first compare the size, compare the size after the change of position, and then continue to change the position, a string of numbers can finally be put out in order.
From 2020 to this year, teachers will say once a year that "artificial intelligence major is good for employment", and this major represents the future direction. "But what is a good job? What is the specific direction? What position? No teacher will say." Zhang Tian still doesn't know.
The artificial intelligence major is regarded as one of the hottest majors at the moment. Every year, colleges and universities will eliminate some outdated and low-employment majors from their faculties, and add some new majors. In 2018, after 35 universities were the first to obtain the qualification for artificial intelligence professional construction, the boom of university construction artificial intelligence major kicked off.
In just four or five years, the artificial intelligence profession has grown rapidly. Up to now, 498 domestic universities have successfully applied for artificial intelligence majors. The boom has not only spread in undergraduate colleges, but also in vocational colleges, the application of artificial intelligence technology has added 71 new professional points last year.
According to the "New Generation Artificial Intelligence Development Plan", by 2025, the scale of the core industry of Chinese industrial intelligence will exceed 400 billion yuan, driving the scale of related industries to exceed 5 trillion yuan. AI students are naturally seen as the reserve army of the emerging AI industry in the future. With the popularity of ChatGPT this year, artificial intelligence majors have attracted much attention.
Colleges and universities are rushing to explore artificial intelligence majors, but at the same time, many ordinary college artificial intelligence students are feeling lost.
In her sophomore year, Li Wanni, who studied artificial intelligence at a university in Shenyang, submitted her internship resume, but only received an invitation to a written test, and finally did not enter the interview. In the junior year, the school organized a double election meeting, and most companies recruited artificial intelligence-related positions with a salary of only 4,000-5,000 yuan - which is far from the imagined high salary. Even so, the image recognition engineer post cast by Li Wanni did not hear back. In one interview, the company asked Li Wanni to know the graph algorithms of "convolution, sharpening and deformity", but she only learned convolution "superficially" in class. "I followed the teacher step by step, but when I look for a job, I will find that what the company wants you to do is not the same thing as the teacher. But where am I going to learn those things?"
At 211 University, a liberal arts university in Beijing, Wang Xiaofen, who was transferred to an artificial intelligence major, has been working hard to change majors since her freshman year, although her family believes that it is the right choice by mistake. She is reluctant to talk to anyone about AI. After the ChatGPT fire this year, relatives at home talked to Wang Xiaofen about the development of AI, and she could only deal with it hastily. In fact, she is still very vague about what "artificial intelligence" is and what it is to do, and she is still very vague in her junior year.
The salary level of artificial intelligence-related positions displayed on a recruitment website is far from the high salary Li Wanni imagined.
A gathering place for "dispensers"
Li Wanni was born in a medical family, her parents are engaged in medicine-related work, after the college entrance examination, her first choice is traditional Chinese medicine, and her second choice is pharmacy. At that time, her long-term plan was to become a pharmacist. What the third volunteer filled in was insignificant in her eyes at that time. My father saw that the university in Shenyang had newly opened a "artificial intelligence major", and in the publicity recommendation, the school explained that this major would cooperate with Huawei. My father said, "Then fill this in, it looks good."
Although Li Wanni had no understanding of this major, she thought that she would not slip to the third choice of admission, so she simply filled in this major. When it came to admission, in the case of pharmacy pressing, she was coincidentally sent to the artificial intelligence major and learned artificial intelligence in vague.
Ke Beilin, a junior major in artificial intelligence at a business university in Hunan Province, is in a similar situation. Her first choice is accounting, international trade and other majors, because she likes mathematics, she also filled in the mathematics major. That year, her university enrolled 4 classes of 40 AI students, with a total of 160 majors. In the case of insufficient enrollment scale, Ke Beilin, who received the transfer, was transferred to the artificial intelligence major after not being admitted to each major. Before the start of school, the topic of transfer in the freshman group was discussed, some people said that they were transferred, and soon someone picked up, "Me too". After the start of school, Ke Beilin learned that there were 4 people in the dormitory, and 3 people were transferred.
Although according to the "Baidu Hot Search 2022 College Entrance Examination Big Data", as of last year, artificial intelligence has ranked as the profession with the highest increase in popularity for three consecutive years. But three years ago, this new major was not in demand in most ordinary universities. At that time, many students and parents knew almost nothing about this new major.
When Wang Xiaofen graduated from high school, she looked at a 211 university in Beijing with liberal arts, and wanted to study media majors, the first few volunteers filled in public relations, network and new media, and artificial intelligence majors were the last volunteers, but she never expected that her score was recorded in artificial intelligence.
She remembers the sense of gap. When she opened the admission page, she was very happy to see that she had gone to her favorite university, and then found that she was assigned to the artificial intelligence major, and the college she belonged to was the "School of Information and Communication Engineering". She couldn't believe it, because she didn't even know that this liberal arts college had this college and major. When filling out the college entrance examination, in order not to continue studying physics and mathematics in college, she also "struggled" with her parents for a long time, but in the end, she still studied engineering.
After enrollment, six people in the dormitory, including Wang Xiaofen, four students were transferred to the artificial intelligence major, and one student mistakenly wrote the artificial intelligence major with a lower score in front because of an operation error.
Zhang Tian is one of the few who puts the artificial intelligence major in the first and second volunteers. Her first choice was computer, and her second choice was artificial intelligence. In high school, a mathematics teacher who graduated from Xi'an Jiaotong University mentioned the artificial intelligence major, saying that "this thing is relatively new, the development may be better, and the employment is quite good." At that time, the big factory dividends were still flowing, Zhang Tian had no ambitions, she felt that as long as she graduated, she could find a job, and she simply believed that computer and artificial intelligence majors should be promising.
Due to various coincidences, high school students who have not yet had any understanding of artificial intelligence have entered a new major. It represents the future, but also the unknown.
A university in Zhengzhou has set up an artificial intelligence major, and students take classes in the AI laboratory. Source: Visual China
"Learn computers under a different name"
What kind of profession is artificial intelligence? Every student who chooses it has to face this problem first-hand.
There is no reference answer. For most students in ordinary colleges and universities, this is almost a major without seniors. What classes to take, which textbooks to use, how to find an internship, which company to go to for an internship, whether to graduate school or employment, these questions are not ready-made, all rely on their own exploration. They themselves are the most "old" students majoring in artificial intelligence.
For schools, how to open an artificial intelligence major, starting from the setting of the name of the college and major, is full of "suspense". At Hunan Technology and Business University, the artificial intelligence major was originally opened under the School of Computer Science, in the same compound as the information science and information management majors. By the second year, the artificial intelligence major was spun off from the School of Computer Science and transferred to a new college: the Frontier Interdisciplinary College. Ke Bellin said that the "intelligent science and technology" major, which is more related to the artificial intelligence major, was separated into the College of Intelligent Manufacturing and Science and Technology.
What puzzled Ke Beilin was that after the artificial intelligence major was divided into the frontier interdisciplinary college, the traditional engineering management was also assigned to this college. She didn't understand the logic of this, but just thought, "It's a bit messy."
At Wang Xiaofen's 211 university, which is dominated by liberal arts, the artificial intelligence major is located under the School of Information and Communication Engineering, not the School of Computer Science set up by most schools. Wang Xiaofen explained, "Because the [school's] ICT College has a higher rating, it has the qualification to open another major."
According to the National Standards for Subject Classification and Code of the People's Republic of China, artificial intelligence is currently one of the secondary disciplines under "Computer Science and Technology". Or for this reason, most undergraduate colleges choose to locate artificial intelligence under the School of Computer Science, or regard it as a major extended of the computer major.
Therefore, the courses of artificial intelligence majors in many ordinary colleges and universities are also very similar to the courses of computer majors. In the first year, almost all artificial intelligence students will learn basic programming languages, and in the sophomore year, they will begin to contact web design, front-end, data structures and other courses, and it is not until the junior year that they begin to contact neural networks, deep learning, intelligent speech, image recognition and other courses with artificial intelligence majors.
A high school classmate asked Li Wanni: "What major did you study?" Li Wanni would even tell them directly, "computer major", "because there is really no difference."
In addition, there are many "special courses" superimposed on the curriculum. In the School of Information and Communication Engineering, Wang Xiaofen takes compulsory courses such as analog circuits, which are not set up in artificial intelligence majors in other schools. The students questioned the teacher, "This class is meaningless", but the teacher shook his head and said, "The teaching plan has been decided and cannot be changed." By the time the next AI major starts, the course has become an elective. Zhang Tian hung up a physics experiment class in her sophomore year. At her school, physics experiments are mandatory for science and engineering.
When there was a lot of schoolwork, Ke Beilin dozed off while studying at the computer. Photo courtesy of interviewee
Because it covers knowledge and technology in multiple disciplines, the coursework of artificial intelligence is more heavy than many people think. Wang Xiaofen recalled her busiest semester, from Monday to Thursday from 8 a.m. to 10 p.m., elective or experimental classes in the evening, and 8 a.m. to 3 p.m. on Fridays. Exhausted, Wang Xiaofen gave up club and department activities, and stayed in the dormitory on weekends to lie down and play with her mobile phone. At that time, internships were not considered at all, because for Wang Xiaofen, who even had few breaks, she could not meet the minimum requirement of most companies to require a three-day on-the-job internship.
A wide variety of courses often take up a full week. She wasn't interested in a course like C, but she listened carefully, practiced, and changed the code like a real programmer until one or two in the morning. But when the subsequent classes became more and more difficult, she simply "attended classes efficiently". Listen to a few lessons first, and give up if you really don't understand.
Since the code foundation has only been studied for more than a year, by the time artificial intelligence courses such as neural networks, pattern recognition, and deep learning really appear, many students can no longer keep up.
Compared with the highly competitive computer major, artificial intelligence students may not be superior. Ke Beilin found that the programs written by computer science seniors are much more mature and complex than artificial intelligence students, "we are still a little skinny, miscellaneous and not refined." Wang Xiaofen also said, "Anyone who talks with computer science students alone, or even looks at their homework, will find that we will not do any practical applications, but simply talk about theory and methods."
In her sophomore year, Wang Xiaofen joined a team internship by finding acquaintances. The senior who brought her told Wang Xiaofen that the team only recruits two kinds of talents, one who can write code and the other who understands AI. The senior also said, "A lot of what is taught in school and what is now developing artificial intelligence is derailed."
After ChatGPT became popular, none of Li Wanni's professional teachers mentioned this tool. At Wang's 211 university, students in AI classes are not allowed to use ChatGPT to complete assignments.
Teachers teach as they learn
In the face of artificial intelligence majors, it is not only students who are not comfortable coping, but also teachers.
For most of the double non-undergraduate universities following the first batch of 35 universities, in addition to the problems of ideas and design, the lack of resources is also obvious.
In ordinary colleges, there are not many teachers with a professional research background in artificial intelligence, and many professional courses are taught by computer teachers who are regarded as an extension of the computer major.
Li Wanni remembers that a teacher of an artificial intelligence major once asked students to study by themselves in class, "You learn by yourself, I may not learn as well as you", and the students laughed. The teacher explained that he urgently participated in about half a month of training during the summer vacation, which was organized by a technology company, and he also took MOOCs, and he learned it once before teaching the content to students.
In the whole hospital, the only teacher Li Wanni learned of with an artificial intelligence background was the young vice dean of the college, a postdoctoral fellow who returned from studying in the United States, and the others were basically teachers with computer background.
Zhang Tian is the kind of person who has worked hard to take every class since her freshman year, but if she sees that the teacher is just reading the PPT on stage with her own care, she knows that this class is most likely "unnecessary".
Guo Zhiwei, associate professor of the School of Information Technology and Information Science of Chongqing Technology and Business University, is one of the teachers involved in the construction of artificial intelligence in the school. During his undergraduate studies, he studied communication engineering at Zhengzhou University and completed his master's and doctoral studies in communication engineering at Chongqing University, specifically studying data mining, which has a certain relationship with artificial intelligence.
He does not deny the challenges faced by the AI profession in the establishment phase. He joined the school in 2018, and before that, the college had submitted materials for applying for artificial intelligence. In 2019, his Institute of Jiaxin began to set up an artificial intelligence major and asked him to prepare to teach "Deep Learning", "because most of the colleges and universities that opened this major at the same time are also starting to practice for the first time, and there is not much experience to learn from each other, and it is actually difficult to do in-depth discussion and teaching (within the college)."
For Guo, deep learning is also new knowledge. Before 2018, he had participated in research in the field of machine learning, "but this is relatively low-level learning, while deep learning is relatively high-level."
There are many difficulties, and high-level knowledge must be learned on your own initiative. Guo Zhiwei participated in the training organized by Baidu, which lasted less than a month, and later participated in various training organized by Tsinghua University Press and other institutions to learn deep learning. He studied for most of the year himself, and in the first half of 2021, he began to teach the course of "Deep Learning". Because he already has some theoretical foundation in artificial intelligence, it is easier to learn than a teacher without a foundation. For him, the most difficult learning is programming like undergraduates, "deep learning models are prone to all kinds of errors, and finding errors how to solve them is also a challenge."
Source: Visual China
The in-depth program is for junior undergraduate students. Before teaching classes in 2021, Guo Zhiwei was already mentally prepared, "My goal is not too high, for undergraduates, many classes do not necessarily take much time, so just master the basic image recognition and text recognition, the real content involving programming is still more difficult."
Because the major is too new, Guo Zhiwei found that "Deep Learning" does not have systematic and standard textbooks, and some existing textbooks are theoretical and lagging behind development, so he has to make his own lectures. This continues into this year, and he is even preparing to write a textbook of his own.
At the same time, artificial intelligence majors need strong enough hardware resources for students to participate in practical courses. Guo Zhiwei explained that deep learning needs to be implemented by programming and programming, because the processing objects are complex and the amount of computing is huge, so personal computers generally cannot carry such a volume. Among them, the key hardware is the GPU, that is, the image processing unit, and the purchase price of a GPU is about 100,000 yuan, which is not a small amount for the college. Although Chongqing Technology and Business University has successively purchased several GPUs since the opening of the major in 2021, it is still "far from meeting the use of most students, only 10-15 personal use, graduate students are not enough points".
Under the restrictions, most of the computer experimental courses of Chongqing Technology and Business University use the cloud computing resources provided by a large domestic factory for teaching and practice. But Guo Zhiwei also said that there are still operational differences between the deep learning tools of this large factory and the mainstream deep learning tools in the world.
For such a rapidly changing discipline as artificial intelligence, it is not easy for teachers to catch up. Taking the deep learning tools of Baidu, a large domestic manufacturer, as an example, Guo Zhiwei used version 2.0 when he first taught in 2021, and this year it is already version 2.4, and he is still exploring and learning new knowledge and tools.
After ChatGPT exploded this year, Guo Zhiwei only mentioned it in class, but did not talk about it in depth. He himself is a little contradictory, is the class easier or harder? And because behind this hot technology is a high-level model, the teacher himself does not know very well.
In his third year of teaching, Kwok was able to sense how undergraduates felt about the major. He said frankly, "I think it is a bit difficult for them to major in artificial intelligence, and their interest has indeed declined in the middle and late stage of learning."
Course content photographed by Ke Beilin Photo courtesy of interviewee
Something that artificial intelligence does not know
When she was most confused, Li Wanni posted on the Internet and asked: "What are the graduation directions to learn, I don't want to put it on the ground."
A graduate student and practitioner in the field of artificial intelligence "strongly" suggested that she change direction, "this direction is not a top education, it is better not to cultivate, but to develop better employment and development." Some people followed behind and said, "Strong ability can go to 985 and strong 211, and artificial intelligence requires a particularly good mathematical foundation." And to engage in algorithm positions, even if you graduate from a 985 bachelor's degree, it is difficult to find a job, and you have to continue to study for a master's degree to have the opportunity to enter the "roll".
Guo Zhiwei told "People" that the reason why most double non-ordinary undergraduates classify artificial intelligence majors as a major evolved from computer majors is also to ensure employment, "Artificial intelligence employment is good, and the school is not bottom, so you must master some computer content, at least do some software development, so you won't find a job." He believes that it is indeed difficult for the undergraduate classroom to fully meet the requirements of actual employment. Some of the students he has brought have also entered the field of artificial intelligence, such as going to large factories to do data mining, "but the proportion is not high, only a few." Today, about 50 to 60 percent of the students in his class choose graduate school.
Li Wanni also remembers that a vice dean of the college directly said, "Artificial intelligence major is to read master's and doctoral degrees, otherwise the knowledge learned is stuck in a very shallow stage." And she has lost confidence in continuing to study artificial intelligence, let alone entering graduate school. The almost unimaginable choice was that she would be part of the exam.
Reality continues to corroborate these claims. Ke Bellin submitted many algorithm-related internship resumes, but they were all rejected. She mentioned that there is an "AI intelligent trainer" in the market, responsible for labeling voice, video, and pictures, and high school students can also do it. Because the job is difficult to find, she has lowered her expectations to the point that this kind of machine is actually only a position with the word "AI", but there is still little response. She had the idea of entering graduate school, but she felt that in order to improve her education, she had to go to 985 and 211, which was "too rolled", and she might not be able to get in.
At the moment, artificial intelligence students are not much different from other undergraduate students. Zhang Tian, who is a sophomore, decided to "follow the flow", join the postgraduate examination team early, and transfer to the computer major. She regretted learning artificial intelligence, but back 3 years ago, "maybe there is no major that can make her choose without regrets."
As the first batch of undergraduates to step into artificial intelligence, although sometimes tired and confused, there is still a chance to touch the amazing side of this technology first. Wang Xiaofen first used ChatGPT during her internship, when the language model was not yet popular and not as smart as it is now. She is responsible for asking ChatGPT questions on the team, constantly adjusting the words and keywords of the questions to get valuable answers. She suddenly felt that "people are still irreplaceable". She believes that even if she interdisciplinary in the future, this artificial intelligence learning experience may be beneficial to her in the future.
For teachers, although teaching work is still difficult, early practitioners like Guo Zhiwei feel that it still makes sense to open artificial intelligence majors at the undergraduate level, because if you really want to cultivate AI talents, you must start from the basic undergraduate stage, but it still takes time to explore the curriculum and learning standards. "This is the reform content to be considered."
Approaching summer, the popularity of artificial intelligence majors continues. The first batch of undergraduate students majoring in artificial intelligence in ordinary colleges and universities are about to graduate. Whether the road ahead is easy to go, AI will not answer, time will be.
Source: Visual China
(At the request of interviewees, Zhang Tian, Li Wanni, Wang Xiaofen, and Ke Beilin are pseudonyms in the article)