Xueba's life simulator
Chapter 196 Doing research is not easy
Chapter 196 Doing research is not easy
"I don't know what this research is that Teacher Bi is talking about? Can you tell me more specifically?" Zhou Ming also asked Bi Wanfang with some curiosity.
After hearing Zhou Ming’s question, Bi Wanfang smiled mysteriously, and said to Zhou Ming: “Actually, it’s related to your previous article [About a neural network and execution system used to train neural networks and Some of the content mentioned in Algorithm Discussion of Other Tasks] is related to the research on deep neural network models.”
Deep neural network, referred to as DNN in English, is good at visual recognition tasks. At the beginning, scientists still used mice for experiments, such as guiding the auditory area of the brain of mice to produce functions similar to vision. OK.
The results of this study can be used to study the installation of "electronic eyes" on blind people, adding bat-like sonar systems to blind people, and then guiding their auditory brain areas to see objects using a method similar to processing visual signals.
The purpose of all scientific experiments is to serve humans, and the differences between mice and humans are too huge, so some scientists have begun to use deep neural networks as a modeling framework for neural computing in primate brains. .
Just like a single brain, each deep neural network has a unique connectivity and representation profile.
"Since the publication of your paper, the development of the field of artificial intelligence compared with the development of other fields of computer is almost like a speed between ordinary express and high-speed rail.
And deep neural networks have now become the focus of the computational neuroscience community. After revolutionizing computer vision with unprecedented task performance earlier this year, the corresponding networks were quickly tested. They explain information processing in the brain. Ability.
My goal this time is to build the best model class for predicting the activities of multiple regions of primate visual cortex through task-optimized deep neural networks.
But even if this visual model were to be constructed, the question remains of how generalized are network internal representations, or whether network instances, like brains, exhibit individual differences due to their unique connectivity characteristics. "
Bi Wanfang continued to talk to Zhou Ming about his research field of computational neuroscience and the problems he might encounter. Even a layman can hear it from what Bi Wanfang said now. His research involves at least two fields of biology and artificial intelligence.
"Your worry is necessary. I think there will definitely be individual differences, and the differences will not be small."
Zhou Ming told Bi Wanfang that although he could be sure about some things, he couldn't express himself with certainty.
So before Zhou Ming expresses some of his own views that have not yet been verified, he always adds a word like I think, I think.
At this moment, the state of chatting between the two is completely different from just now. Just now they were chatting about things, so they were more casual.
But now they are talking about scientific research, and scientific research is rigorous, so they naturally don't have the casualness just now.
"The large variance means that the common practice of analyzing individual network instances is wrong, and that groups of networks need to be analyzed to ensure the validity of the insights gained. So we then also investigate individual differences between deep neural network instances and investigate this research……"
Bi Wanfang followed Zhou Ming's words and continued.
The two chatted on this topic for a long time, and Zhou Ming also put forward many of his own opinions. After the two finally chatted, Bi Wanfang also appeared very happy. He said to Zhou Ming: "It seems that I invited you to join in." This move can be regarded as the most correct move I made."
However, Zhou Ming did not say any more polite words to Bi Wanfang, but asked Bi Wanfang another question: "This national engineering laboratory for brain-like intelligence technology and application, although the main body of the construction is Huaguo University of Science and Technology, But it should have united with some other units, right? Like other universities, companies and the like.”
"Yes, it seems that you also understand this aspect very well. This laboratory has indeed combined the domestic superior scientific research and industrial resources in the field of brain-inspired intelligence, because only in this way can a joint construction group of universities, scientific research institutes, and high-tech enterprises be formed.
As for the joint construction units of the laboratory, including Fudan University in Shanghai, Shengjing Institute of Automation of Huaguo Academy of Sciences, Institute of Microelectronics of Huaguo Academy of Sciences, Baidu Company, Institute of Electronics of Huaguo Academy of Sciences, and Neuroscience Research of Huaguo Academy of Sciences Institute, HKUST Xunfei Co., Ltd..."
Bi Wanfang finally said a long list of company names. Zhou Ming frowned after listening to it. No wonder he didn’t know about it. It turned out that the joint construction unit of the National Engineering Laboratory for Brain-Inspired Intelligence Technology and Application has no future. The name of the smart technology company.
But think about it, the National Engineering Laboratory for Brain-inspired Intelligent Technology and Applications was established at the same time, and Zhou Ming’s Future Intelligent Technology Company was established less than a year ago. The most outstanding achievement of this year is the ability to accurately predict Artificial intelligence for protein structure.
Although there have been some other achievements since then, none of them can match this artificial intelligence that can accurately predict protein structures. Such an artificial intelligence is now in the leading position in the country, and it is also the only national scientific research platform in the field of brain-like intelligence in China. Naturally, it would not be possible to bring in a future artificial intelligence technology company that requires qualifications but no qualifications, and requires research results but has little results.
"What's the matter?" Bi Wanfang asked Zhou Ming strangely after seeing that the conversation was almost done, but Zhou Ming fell into deep thought.
"Oh? It's nothing. I thought of something. My Future Intelligent Technology Company is also doing research in this area." Since this laboratory did not bring Future Intelligent Technology Company in, Zhou Ming did not want to let Future Intelligent Technology Company forcefully Squeeze in.
"Really? I haven't heard of it before, but I heard that your company and Xunfei have invested a lot of initial funds in order to develop the self-developed top-level chip, and still have money for artificial intelligence?"
I heard Zhou Ming say that future smart technology companies are also conducting research on artificial intelligence. Although this is indeed a reasonable thing, it still surprised Bi Wanfang.
It is reasonable to say that it is because the future smart technology company started with artificial intelligence, and with Zhou Ming, an expert in the field of artificial intelligence, as the boss, it is strange that it does not research artificial intelligence.
The reason why Bi Wanfang was surprised was that in his view, or in other words, in the eyes of most people, Future Smart Technology Company, a small company that has just been established, does not have many profitable projects. All the money is invested in the research of chips, where is the money to study artificial intelligence?
Although artificial intelligence has a wide range of applications, whether it is manufacturing or finance, or agriculture and scientific research, artificial intelligence can be of great use.
However, at the same time, artificial intelligence is also recognized as a field that is both expensive and labor-intensive. If you want to really do it, its money-burning ability is no less than that of independently developing top-level chips.
Many people even suspect that the money used by Future Smart Technology Company to research chips is likely to be invested by another company of its boss, Future Biotechnology Company, and that Future Biotechnology Company also needs a lot of money to develop itself.
Where do smart technology companies in the future get so much energy and money to develop side by side?
"Hehe, multi-line development." Facing Bi Wanfang's question, Zhou Ming just smiled and didn't say much.
By this time, Bi Wanfang had finally seen Zhou Ming's ambitions in this field.
If you only study artificial intelligence, even if you research it beyond the artificial intelligence of any technology company on the planet, it is useless without a corresponding carrier.
Just like human consciousness, where does consciousness exist without the brain?
In Bi Wanfang's view, Zhou Ming's purpose of developing multiple lines side by side is to prevent this kind of thing from happening.
However, this is obviously an extremely daring operation, because once the intermediate capital chain is broken, you have no money to continue, and there are no research results, it will not only be as simple as nothing.
"Zhou Ming seems to have always been so prepared for a rainy day?" Bi Wanfang thought of some thoughts he had said when he and Zhou Ming were talking about some things before.
In the eyes of some people, Zhou Ming's behavior is called planning for a rainy day, but in the eyes of others, it is called unfounded worry.
In the end, Bi Wanfang just looked at Zhou Ming and said nothing more.
After all, Zhou Ming has been creating the impossible since his rise last year, no matter whether he has acquired dozens or even tens of billions of future biotechnology companies within a year of his establishment, or proved the twin prime number conjecture and Goldbach conjecture successively. , the research of super beans, industrialized artificial synthesis of glucose, starch and other achievements, in the eyes of people in the past, it was impossible for the same person to complete them.
Now, who knows whether this time will be just one of the countless impossibilities that Zhou Ming will create in the future?
Not long after, Zhou Ming left Bi Wanfang's office. It had been more than three hours since he came.
Zhou Ming went to the cafeteria to have a meal, and then returned to his office to prepare for a break. Zhou Ming just squinted his eyes when he heard a "ding dong" notification sound from his mobile phone.
Zhou Ming grabbed his phone from the table, opened his eyes, unlocked the screen and saw that it was a message from WeChat.
The person who sent the message was none other than Shen Qingqiu's mentor, Bi Wanfang's wife, Chen Yanhong, a professor at the Department of Genetics and Cell Biology at Nankai University.
After Zhou Ming added Professor Chen as a friend, Chen Yanhong sent a series of text messages.
"Hello, Professor Zhou, I heard from Wan Fang that you are going to recommend a research project as student Shen's master's graduation project. I wonder if you can talk about this research in detail?"
Seeing the string of text sent by Chen Yanhong, Zhou Ming also sent a message.
"Professor Chen, let's make a voice call to chat? Or I can call you directly."
For this kind of academic matter, it is too cumbersome to send text, and it is too time-consuming to send voice. After you finish speaking and send it, you have to ask the other party to listen to it again before she can send you a message.
And after the other party sends it, you have to listen to it again before sending the message.
It takes a lot of time to go back and forth.
A few seconds after Zhou Ming's message was sent, his mobile phone rang a WeChat voice call, and it was Chen Yanhong who called.
Then, Zhou Ming chatted with Chen Yanhong about the molecular mechanism of endoplasmic reticulum proteins regulating the distribution of organelles that he was going to ask Shen Qingqiu to conduct.
"...Well, yes, the first thing to explore is the function of these three membrane-bound endoplasmic reticulum proteins in regulating the shape and distribution of the endoplasmic reticulum...
Another point is that in order to better quantitatively evaluate the changes in ER morphology and distribution, you can use the complementary algorithm I created previously, and use statistical methods based on probability density estimation to analyze the spatial distribution of fluorescently labeled ER and other organelles. The resulting spatial probability mass function was used to quantify the fluorescence change on the image to calculate the radial distribution of organelles and the degree of cellular asymmetry…
Well, definitely something to do with the tubulin code in microtubule dynamics and information encoding..."
It took Zhou Ming nearly an hour to roughly explain to Chen Yanhong the general idea of the research he was going to let Shen Qingqiu conduct, some other research involved in it, and the research results he predicted.
After understanding Zhou Ming's idea, Chen Yanhong was a little hesitant, and finally she said to Zhou Ming: "Although this research result comes out, Student Shen will definitely become a blockbuster and be known by people because of this research result.
But even if this research has your remote guidance and my on-site guidance, as well as some of the other researches handed over to my other graduate students, Shen Qingqiu is probably not an easy task to complete it alone.
Are you sure you want Mr. Shen to do this research? "
It turned out that Chen Yanhong was worried for Shen Qingqiu, because as far as what Zhou Ming said, it was not something a rookie in scientific research could handle alone. Even if the two professors knew about it, it would be very difficult.
It has very high requirements on researchers' knowledge reserve, scientific research experience, hands-on ability and other abilities, and not everyone is a pervert like Zhou Ming.
After hearing what Professor Chen said, Zhou Ming finally realized that there may indeed be many things that seem very simple to him, but may be very difficult to Shen Qingqiu.
Therefore, Zhou Ming also said to Chen Yanhong: "I really didn't think of this before. Thank you Professor Chen for reminding me. Then I will ask Qingqiu and see what she thinks before making a decision."
After hanging up the phone with Zhou Ming, Professor Chen Yanhong, who was far away in Nankai, also felt a little speechless. This Zhou Ming's scientific research ability was incredible, but he was too young, had no experience in some things, and his thinking was too simple.
This kind of thing, unexpectedly ran to him without explaining it clearly to Shen Qingqiu.
But Zhou Ming looked at the time after hanging up the phone, and found that it would be Shen Qingqiu's class time soon.
If you want to clarify this matter with Shen Qingqiu, you can't just say a few words, so Zhou Ming didn't call Shen Qingqiu at this time, but planned to talk to her when Shen Qingqiu had enough time at night.
This chapter is for today.
(End of this chapter)
"I don't know what this research is that Teacher Bi is talking about? Can you tell me more specifically?" Zhou Ming also asked Bi Wanfang with some curiosity.
After hearing Zhou Ming’s question, Bi Wanfang smiled mysteriously, and said to Zhou Ming: “Actually, it’s related to your previous article [About a neural network and execution system used to train neural networks and Some of the content mentioned in Algorithm Discussion of Other Tasks] is related to the research on deep neural network models.”
Deep neural network, referred to as DNN in English, is good at visual recognition tasks. At the beginning, scientists still used mice for experiments, such as guiding the auditory area of the brain of mice to produce functions similar to vision. OK.
The results of this study can be used to study the installation of "electronic eyes" on blind people, adding bat-like sonar systems to blind people, and then guiding their auditory brain areas to see objects using a method similar to processing visual signals.
The purpose of all scientific experiments is to serve humans, and the differences between mice and humans are too huge, so some scientists have begun to use deep neural networks as a modeling framework for neural computing in primate brains. .
Just like a single brain, each deep neural network has a unique connectivity and representation profile.
"Since the publication of your paper, the development of the field of artificial intelligence compared with the development of other fields of computer is almost like a speed between ordinary express and high-speed rail.
And deep neural networks have now become the focus of the computational neuroscience community. After revolutionizing computer vision with unprecedented task performance earlier this year, the corresponding networks were quickly tested. They explain information processing in the brain. Ability.
My goal this time is to build the best model class for predicting the activities of multiple regions of primate visual cortex through task-optimized deep neural networks.
But even if this visual model were to be constructed, the question remains of how generalized are network internal representations, or whether network instances, like brains, exhibit individual differences due to their unique connectivity characteristics. "
Bi Wanfang continued to talk to Zhou Ming about his research field of computational neuroscience and the problems he might encounter. Even a layman can hear it from what Bi Wanfang said now. His research involves at least two fields of biology and artificial intelligence.
"Your worry is necessary. I think there will definitely be individual differences, and the differences will not be small."
Zhou Ming told Bi Wanfang that although he could be sure about some things, he couldn't express himself with certainty.
So before Zhou Ming expresses some of his own views that have not yet been verified, he always adds a word like I think, I think.
At this moment, the state of chatting between the two is completely different from just now. Just now they were chatting about things, so they were more casual.
But now they are talking about scientific research, and scientific research is rigorous, so they naturally don't have the casualness just now.
"The large variance means that the common practice of analyzing individual network instances is wrong, and that groups of networks need to be analyzed to ensure the validity of the insights gained. So we then also investigate individual differences between deep neural network instances and investigate this research……"
Bi Wanfang followed Zhou Ming's words and continued.
The two chatted on this topic for a long time, and Zhou Ming also put forward many of his own opinions. After the two finally chatted, Bi Wanfang also appeared very happy. He said to Zhou Ming: "It seems that I invited you to join in." This move can be regarded as the most correct move I made."
However, Zhou Ming did not say any more polite words to Bi Wanfang, but asked Bi Wanfang another question: "This national engineering laboratory for brain-like intelligence technology and application, although the main body of the construction is Huaguo University of Science and Technology, But it should have united with some other units, right? Like other universities, companies and the like.”
"Yes, it seems that you also understand this aspect very well. This laboratory has indeed combined the domestic superior scientific research and industrial resources in the field of brain-inspired intelligence, because only in this way can a joint construction group of universities, scientific research institutes, and high-tech enterprises be formed.
As for the joint construction units of the laboratory, including Fudan University in Shanghai, Shengjing Institute of Automation of Huaguo Academy of Sciences, Institute of Microelectronics of Huaguo Academy of Sciences, Baidu Company, Institute of Electronics of Huaguo Academy of Sciences, and Neuroscience Research of Huaguo Academy of Sciences Institute, HKUST Xunfei Co., Ltd..."
Bi Wanfang finally said a long list of company names. Zhou Ming frowned after listening to it. No wonder he didn’t know about it. It turned out that the joint construction unit of the National Engineering Laboratory for Brain-Inspired Intelligence Technology and Application has no future. The name of the smart technology company.
But think about it, the National Engineering Laboratory for Brain-inspired Intelligent Technology and Applications was established at the same time, and Zhou Ming’s Future Intelligent Technology Company was established less than a year ago. The most outstanding achievement of this year is the ability to accurately predict Artificial intelligence for protein structure.
Although there have been some other achievements since then, none of them can match this artificial intelligence that can accurately predict protein structures. Such an artificial intelligence is now in the leading position in the country, and it is also the only national scientific research platform in the field of brain-like intelligence in China. Naturally, it would not be possible to bring in a future artificial intelligence technology company that requires qualifications but no qualifications, and requires research results but has little results.
"What's the matter?" Bi Wanfang asked Zhou Ming strangely after seeing that the conversation was almost done, but Zhou Ming fell into deep thought.
"Oh? It's nothing. I thought of something. My Future Intelligent Technology Company is also doing research in this area." Since this laboratory did not bring Future Intelligent Technology Company in, Zhou Ming did not want to let Future Intelligent Technology Company forcefully Squeeze in.
"Really? I haven't heard of it before, but I heard that your company and Xunfei have invested a lot of initial funds in order to develop the self-developed top-level chip, and still have money for artificial intelligence?"
I heard Zhou Ming say that future smart technology companies are also conducting research on artificial intelligence. Although this is indeed a reasonable thing, it still surprised Bi Wanfang.
It is reasonable to say that it is because the future smart technology company started with artificial intelligence, and with Zhou Ming, an expert in the field of artificial intelligence, as the boss, it is strange that it does not research artificial intelligence.
The reason why Bi Wanfang was surprised was that in his view, or in other words, in the eyes of most people, Future Smart Technology Company, a small company that has just been established, does not have many profitable projects. All the money is invested in the research of chips, where is the money to study artificial intelligence?
Although artificial intelligence has a wide range of applications, whether it is manufacturing or finance, or agriculture and scientific research, artificial intelligence can be of great use.
However, at the same time, artificial intelligence is also recognized as a field that is both expensive and labor-intensive. If you want to really do it, its money-burning ability is no less than that of independently developing top-level chips.
Many people even suspect that the money used by Future Smart Technology Company to research chips is likely to be invested by another company of its boss, Future Biotechnology Company, and that Future Biotechnology Company also needs a lot of money to develop itself.
Where do smart technology companies in the future get so much energy and money to develop side by side?
"Hehe, multi-line development." Facing Bi Wanfang's question, Zhou Ming just smiled and didn't say much.
By this time, Bi Wanfang had finally seen Zhou Ming's ambitions in this field.
If you only study artificial intelligence, even if you research it beyond the artificial intelligence of any technology company on the planet, it is useless without a corresponding carrier.
Just like human consciousness, where does consciousness exist without the brain?
In Bi Wanfang's view, Zhou Ming's purpose of developing multiple lines side by side is to prevent this kind of thing from happening.
However, this is obviously an extremely daring operation, because once the intermediate capital chain is broken, you have no money to continue, and there are no research results, it will not only be as simple as nothing.
"Zhou Ming seems to have always been so prepared for a rainy day?" Bi Wanfang thought of some thoughts he had said when he and Zhou Ming were talking about some things before.
In the eyes of some people, Zhou Ming's behavior is called planning for a rainy day, but in the eyes of others, it is called unfounded worry.
In the end, Bi Wanfang just looked at Zhou Ming and said nothing more.
After all, Zhou Ming has been creating the impossible since his rise last year, no matter whether he has acquired dozens or even tens of billions of future biotechnology companies within a year of his establishment, or proved the twin prime number conjecture and Goldbach conjecture successively. , the research of super beans, industrialized artificial synthesis of glucose, starch and other achievements, in the eyes of people in the past, it was impossible for the same person to complete them.
Now, who knows whether this time will be just one of the countless impossibilities that Zhou Ming will create in the future?
Not long after, Zhou Ming left Bi Wanfang's office. It had been more than three hours since he came.
Zhou Ming went to the cafeteria to have a meal, and then returned to his office to prepare for a break. Zhou Ming just squinted his eyes when he heard a "ding dong" notification sound from his mobile phone.
Zhou Ming grabbed his phone from the table, opened his eyes, unlocked the screen and saw that it was a message from WeChat.
The person who sent the message was none other than Shen Qingqiu's mentor, Bi Wanfang's wife, Chen Yanhong, a professor at the Department of Genetics and Cell Biology at Nankai University.
After Zhou Ming added Professor Chen as a friend, Chen Yanhong sent a series of text messages.
"Hello, Professor Zhou, I heard from Wan Fang that you are going to recommend a research project as student Shen's master's graduation project. I wonder if you can talk about this research in detail?"
Seeing the string of text sent by Chen Yanhong, Zhou Ming also sent a message.
"Professor Chen, let's make a voice call to chat? Or I can call you directly."
For this kind of academic matter, it is too cumbersome to send text, and it is too time-consuming to send voice. After you finish speaking and send it, you have to ask the other party to listen to it again before she can send you a message.
And after the other party sends it, you have to listen to it again before sending the message.
It takes a lot of time to go back and forth.
A few seconds after Zhou Ming's message was sent, his mobile phone rang a WeChat voice call, and it was Chen Yanhong who called.
Then, Zhou Ming chatted with Chen Yanhong about the molecular mechanism of endoplasmic reticulum proteins regulating the distribution of organelles that he was going to ask Shen Qingqiu to conduct.
"...Well, yes, the first thing to explore is the function of these three membrane-bound endoplasmic reticulum proteins in regulating the shape and distribution of the endoplasmic reticulum...
Another point is that in order to better quantitatively evaluate the changes in ER morphology and distribution, you can use the complementary algorithm I created previously, and use statistical methods based on probability density estimation to analyze the spatial distribution of fluorescently labeled ER and other organelles. The resulting spatial probability mass function was used to quantify the fluorescence change on the image to calculate the radial distribution of organelles and the degree of cellular asymmetry…
Well, definitely something to do with the tubulin code in microtubule dynamics and information encoding..."
It took Zhou Ming nearly an hour to roughly explain to Chen Yanhong the general idea of the research he was going to let Shen Qingqiu conduct, some other research involved in it, and the research results he predicted.
After understanding Zhou Ming's idea, Chen Yanhong was a little hesitant, and finally she said to Zhou Ming: "Although this research result comes out, Student Shen will definitely become a blockbuster and be known by people because of this research result.
But even if this research has your remote guidance and my on-site guidance, as well as some of the other researches handed over to my other graduate students, Shen Qingqiu is probably not an easy task to complete it alone.
Are you sure you want Mr. Shen to do this research? "
It turned out that Chen Yanhong was worried for Shen Qingqiu, because as far as what Zhou Ming said, it was not something a rookie in scientific research could handle alone. Even if the two professors knew about it, it would be very difficult.
It has very high requirements on researchers' knowledge reserve, scientific research experience, hands-on ability and other abilities, and not everyone is a pervert like Zhou Ming.
After hearing what Professor Chen said, Zhou Ming finally realized that there may indeed be many things that seem very simple to him, but may be very difficult to Shen Qingqiu.
Therefore, Zhou Ming also said to Chen Yanhong: "I really didn't think of this before. Thank you Professor Chen for reminding me. Then I will ask Qingqiu and see what she thinks before making a decision."
After hanging up the phone with Zhou Ming, Professor Chen Yanhong, who was far away in Nankai, also felt a little speechless. This Zhou Ming's scientific research ability was incredible, but he was too young, had no experience in some things, and his thinking was too simple.
This kind of thing, unexpectedly ran to him without explaining it clearly to Shen Qingqiu.
But Zhou Ming looked at the time after hanging up the phone, and found that it would be Shen Qingqiu's class time soon.
If you want to clarify this matter with Shen Qingqiu, you can't just say a few words, so Zhou Ming didn't call Shen Qingqiu at this time, but planned to talk to her when Shen Qingqiu had enough time at night.
This chapter is for today.
(End of this chapter)
You'll Also Like
-
The female CEO dumped him for 5 million yuan in breakup fee
Chapter 201 8 hours ago -
Conan: The winery killer has a soft spot for Huiyuan
Chapter 163 8 hours ago -
Surprise! This son of God is actually a villain
Chapter 198 8 hours ago -
The beginning of the extreme emperor soldier, confession of the villain empress
Chapter 184 9 hours ago -
Gao Wu: I transform into Yama of the underworld to judge life and death
Chapter 213 9 hours ago -
Relying on the beauty master to be invincible, do you know how hard it is?
Chapter 114 9 hours ago -
Saint! You don't want to see the fall of the Heavenly Demon Sect, do you?
Chapter 233 9 hours ago -
Live: My Ocean Ranch
Chapter 259 9 hours ago -
Genshin Impact: I am not a grass god
Chapter 239 9 hours ago -
Beast Tamer Simulator: My beast can evolve infinitely
Chapter 223 9 hours ago