He just read the introduction and immediately realized that the paper might be helpful in improving data models and algorithms.
Xigbarn is different from Alvin. He will not study it slowly by himself. Instead, he simply calls a few people over, "Everyone put down what you are doing and take a look at this together first."
At this time, it lasted more than two hours.
They also concluded, “This method of assisting in building data models is indeed feasible.”
"According to this method, the more data, the higher the efficiency and accuracy, but when there is less data, it is not applicable."
"The difficulty lies in the construction of the original model. It is impossible for us to overturn the original structure and reconstruct it using this method."
"The method is new, but it is not easy to apply."
Sigbarn leads the core algorithm group, and several of his employees can be said to be the elites of the R&D center.
After careful study and discussion, they already have a basic understanding of the construction methods mentioned in the paper, and can infer the situation after use.
Finally, Sigbarn concluded, "Generally speaking, for data analysis problems, the more data, the lower the accuracy."
"This model is no exception, but the accuracy decreases very slowly. For example, for one hundred data, the construction model we use has an accuracy rate of 99.9%. If it is 10 billion data, it will
It becomes eighty percent.”
"With an algorithm built using this model, the accuracy rate for one hundred data is only 90%, and for 10 billion data, the accuracy rate will not be lower than 85%."
"This method of model construction is indeed very meaningful, but in terms of application, we still need to study it slowly."
"In the paper, many of them are just general introductions, but the methods are okay. Maybe we can give it a try. If there is a suitable project, we can adopt this method."
…
Data and Data Applications Laboratory, Google R&D Center, San Francisco, California.
Blake Jones is sitting in the office. He has worked at Google for eleven years and has participated in algorithm customization, Android system improvement, artificial intelligence and other projects. He is an extremely experienced and high-level algorithm engineer.
Now he looks at the computer journals he ordered.
There were three papers in a row that attracted his attention. He frowned and read them carefully for a long time. He felt that the methods described in them were very interesting, so he simply took a photo and shared it in the colleague group, "Everyone, take a look at this, it's new.
Data model building methods.”
"I took a look and it was very interesting."
soon.
Several colleagues who had not ordered yet asked Blake-Jones to take some more pictures and send them, preferably all the papers.
Blake-Jones tried the function of extracting text from pictures and found that when it came to mathematical content, many of them were extracted in a mess, so he simply took pictures one by one with a grimace.
This matter was originally in the past.
In the afternoon, the technical manager found Blake-Jones and others and said seriously, "Blake, I also read what you posted. It's very interesting, and more importantly, it may be valuable."
"We have to pay attention."
"Guys, put down what you are doing and study it together. This method may improve our efficiency..."
…
The same thing happened in the R&D centers of several large Internet companies, and even national-level data control centers and information centers.
When a new data model construction method emerges that can be applied to massive data analysis work and improve efficiency and accuracy, it will definitely receive the attention of Internet-related companies and information centers.
Just one day later, news was also received in China.
In the big data R&D center of Aba Cloud Group, Feng Youli, director of the technical department, received a report from his foreign employees--
In a paper published in the new issue of "Computer Mathematics and Information Engineering", a new data model construction method is introduced.
“Deserves attention!”
Special emphasis is placed on this when employees report.
The opening of online ordering and browsing will be delayed for three days. It is not yet available in China, but there are many ways to bypass the delay.
Feng Youli quickly got the entire content of the paper, and he found several high-level technical employees to analyze it carefully together.
At the same time, Feng Youli had an idea, "The author of the paper is a faculty member of Xihai University. He can develop a completely new way of constructing data models. He must be an extremely outstanding talent."
"The simplest and most direct way is to invite him to work in Abayun."
"He is a talent. What Abayun needs most is talent."
"If this method is feasible, related applications, and the correction and improvement of the underlying algorithm will be most perfect if he takes the lead."
…
When the impact of the paper slowly fermented, Wang Hao was not affected for the time being. He and Zhang Zhiqiang completed the research on "Cauchy non-negative matrix factorization expression" together, and the paper was handed over to Zhang Zhiqiang for publication.
Then, he felt at ease.
He sat in the office drinking coffee, chatting with others about daily life, gossip about the school, and even talked about issues between men and women.
The next day was Thursday, and the morning was still very leisurely.
Wang Hao opened the system and took a look at task two. Not surprisingly, he found that the inspiration value was only '33' points. "It seems that research and research are different."
"Even if the research difficulty is the same, the inspiration value acquisition speed is different."
"In research related to Fourier transform, the inspiration value grows faster, while in research on the 'Cauchy problem', the inspiration value grows slower..."
"This probably involves many factors, such as teaching content, students' situation, students' thinking, etc."
Wang Hao sighed lightly and simply started studying the 'Partial Differential Equations'. The inspiration value of the 'Partial Differential Equations' study has exceeded seventy points. At this value, it is almost time to start research.
However, progress is slow.
Two consecutive hours of full dedication can almost be described as "nothing gained", not even a single problem was solved.
"It's difficult!"
"Insufficient inspiration value means that the knowledge and ideas are not enough to support the completion of research. It is too difficult to solve the problem in a hurry!"
He simply changed his job.
Write lesson plans!
Other teachers focus on research and development and writing papers, but for Wang Hao, teaching is the foundation of the foundation. He hopes to continue to improve in teaching, not to study how to make students understand the knowledge content, but to guide students to learn more.
Think more.
This is critical.
When students can think more, they can give back more knowledge and inspiration.
"Guide students to think more..."
"How to guide it?"
"The lecture topics are better. If we are talking about basic knowledge, how can we add content to guide students to think?"
I couldn't think of anything for a while.
Wang Hao inquired with several people in the office, and also asked Zhou Qingyuan, and found out about an old professor--
Wang Huanxin.
Wang Huanxin is fifty-nine years old, but he is still only an associate professor. This is mainly because he has not made any breakthroughs in scientific research. Most of the papers he has published are related to teaching. He is recognized as an outstanding old teacher in the school, and students do not think highly of him.
Difference.
It's just a pity.
No matter how good your teaching is, it is difficult to get promoted to a professional title. He still has not been promoted to a full professor.
People in the office had various gossips, and Zhu Ping said with certainty, "The school is planning to take advantage of the popularity of 'eliminating theoreticalism' and want to give Wang Huanxin a full professor, a teaching-oriented full professor."
This is extremely rare.
Wang Hao knew that Wang Huanxin had strong teaching ability, so he went to attend Wang Huanxin's class in the afternoon. After class, he also found Wang Huanxin to discuss teaching issues.
Wang Huanxin was very enthusiastic about Wang Hao. He said with emotion, "I didn't expect that there would be someone who could talk to me about teaching issues."
"Nowadays, young teachers like you only care about scientific research and papers. However, teachers, the most important thing is to teach students. Teaching students well is the first priority!"
"If you do good scientific research and write good papers, it is better to be a researcher in a laboratory than to be a writer."
Wang Hao nodded in agreement, "I just hope to teach students well and make the courses more efficient."
Wang Huanxin was indeed very happy, and he took the initiative to share his teaching experience.
Wang Hao raised the question of ‘how to guide students to think’, and he also gave several examples to illustrate in detail.
The two of them talked for more than an hour.
It wasn't until he found that Wang Huanxin looked a little tired that Wang Hao thanked him profusely and left. He felt that he had gained a lot and learned several tips on "guiding students to think".
Summarizing these together, you can even write a 'teaching' paper.
However, his thinking is still dominated by logic, and when it comes to making summaries and writing papers, a question immediately comes to his mind: "If you guide students to think more, the teaching effect may not be better."
"For the teaching of basic content, it is enough for students to remember and understand it. So how to judge whether guiding students to think can bring more knowledge and inspiration? Maybe the original teaching method is better?"
This is uncertain.
Wang Hao thought of a way to judge, "If you only give lectures to a few students, the students' IQs are relatively average and it will be difficult to understand the knowledge content."
"Using two methods to compare two knowledge of similar difficulty, maybe we can know which method is better?"
"Then who should we look for? Stupid students, stupid students..."
Wang Hao kept chanting the words "stupid student", and the figures of two girls across from each other suddenly appeared in his mind.