After finishing his morning chat with Zhang Wenduo, Zhao Wenyuan returned to his dean's office.
It must be said that Zhang Wenduo, this newcomer, gave the veteran researcher Zhao Wenyuan a little shock.
They're like a newborn calf that's not afraid of a tiger; they dare to calculate anything.
Predicting the group interaction behavior of billions of individuals—wouldn't it be great if you could Laplace?
Despite his complaints, Zhao Wenyuan couldn't help but ponder: if they were to really research in this direction, what feasible methods were there?
The first step is definitely clustering.
Although there are billions of individuals, the differences between people are not that great when it comes to specific topics.
If we really break it down, it's nothing more than supporting, opposing, or waiting and seeing. There might be more actions that can be taken, but it's still a finite state machine.
Therefore, even if we need to represent the state of 10 billion individuals during the calculation process, we can use vectors of a few orders of magnitude lower.
There's also the issue of the data that Zhang Wenduo mentioned.
The challenges here are different from those faced when training Glutinous Rice Ball. Intent understanding requires high-quality annotations, but group behavior prediction requires the completeness of the signal dimensions. What a person will do is itself a very complex decision-making process, the result of multiple factors working together.
One of the fallacies that humans easily fall into is the Fundamental Attribution Error, and the same applies to models.
For example, a woman comes home from get off work after a long day and finds her husband sitting on the sofa looking gloomy.
She immediately assumed the other person was dissatisfied with her, and the two had a big argument.
But she never expected that the real reason was that her husband's favorite team was defeated in the last minute of the game.
This is the Fundamental Attribution Error; the woman failed to gather the crucial signal of the game's outcome when making her inferences, leading to an incorrect attribution.
Similarly, to train World Model, it is necessary to ensure that the causal pairs in the training data are accurate. The cause may contain multiple dimensions, but in the real world, this kind of data is the most difficult to collect. Often, the data provider itself does not know where it came from or what information led to its behavior.
I wonder if Boss Han has any solutions for this kind of data...
As he pondered this, Zhao Wenyuan suddenly slapped his forehead.
Being led astray by little kids isn't the most pressing issue right now.
Zhao Wenyuan glanced at today's schedule; another visitor was coming soon.
It was Jiang Jiusi, a medical entrepreneur I met by chance at a dinner hosted by Li Yan. He was dissatisfied with the ability of Zhao Wenyuan and Han Luyi to use the Dragon Carving Tangyuan Model, which was not yet available on Seek Him Technology at the time.
This is also the first scenario in which Glutinous Rice Ball have been applied in the Medical industry.
I just don't know what he came here for today.
As if on Cao Cao, Zhao Wenyuan was pondering this when a knock came at the Cao Cao.
"President Zhao, Mr. Jiang from Jianghe Medical has arrived. Shall I go down to greet him?" It was the voice of Fang Fang, Zhao Wenyuan's newly hired secretary.
"Okay, come straight to me after you get on."
A few minutes later, Fang Fang knocked on the door again.
"Please come in," Zhao Wenyuan said loudly as he stood up.
The office door opened, and Jiang Jiusi walked in.
His current state was completely different from when Zhao Wenyuan first met him.
At that time, Jiang Jiusi's company was facing difficulties in financing and its commercialization was not going smoothly. Jiang Jiusi was obviously under a lot of pressure. Even when he went to dinner parties, he didn't take good care of himself and even looked a bit sloppy.
But Jiang Jiusi, who just entered, was radiating confidence and charisma. He was dressed in a high-quality suit, and his glasses were now gold-rimmed; he exuded the air of a successful man. As soon as he entered, he greeted everyone warmly, "President Zhao, Long Time No See! Congratulations!"
Zhao Wenyuan led Jiang Jiusi to the sofa and asked with a smile, "President Jiang is indeed in high spirits, but I wonder if you're congratulating me—"
"I've recently seen many companies online buying up Glutinous Rice Ball's PI (pixels in English) in large quantities. Glutinous Rice Ball has practically become synonymous with domestically produced PIs. Shouldn't we be congratulating ourselves on this?"
Interestingly, many companies make high-profile purchases Glutinous Rice Ball Pi (pi) and even issue statements about it, but it's just a gesture and they may not actually be using it. After all, there are still many challenges to overcome in getting Pi to practical applications.
However, Jiang Jiusi and Jianghe Medical, which he founded, truly reaped the benefits of this Glutinous Rice Ball.
During the early pilot cooperation period, they relied on their strong ability of Tangyuan 0.9 pi to secure digital transformation orders from several hospitals in Beijing, successfully saving themselves from bankruptcy. I heard that they have recently completed a new round of financing.
So although many people don't believe that purchasing Glutinous Rice Ball can accomplish any kind of digital transformation, Jiang Jiusi does believe it, because he himself is a living example.
Zhao Wenyuan chuckled.
The fact that Tangyuan 0.9 was entrusted to Dingsheng Sales was only announced this Monday. This matter is not under the jurisdiction of Zhao Wenyuan's Beijing Research Institute. If he wants to know more data, he will have to wait until Monday's meeting of senior executives of the Origin Group.
"Tea? Coffee?" Zhao Wenyuan asked.
Jiang Jiusi glanced around Zhao Wenyuan's office and saw that there wasn't even a tea table. What was he supposed to drink tea in there?
"A cup of coffee, please," Jiang Jiusi said. "Two cups of coffee," Zhao Wenyuan said to Fang Fang, who was still waiting at the door, then turned to Jiang Jiusi and asked, "Mr. Jiang, what brings you here today that you need my guidance on something?"
"President Zhao. President Zhao, you flatter me. We owe our success today entirely to you," Jiang Jiusi said, adjusting his glasses. "To be honest, I'm here today to ask you—and President Zhao. Boss Han —for a favor. But it's not for me, it's for a friend of mine."
Zhao Wenyuan raised an eyebrow.
How is Boss Han involved here?
To be honest, the cooperation with Jiang Jiusi was all promoted by Zhao Wenyuan in Beijing, so the two of them were naturally more familiar with each other. It's not hard to understand why Jiang Jiusi would come to him instead of going directly to Boss Han.
The problem is, in the Beijing, is there anything Zhao Wenyuan absolutely incapable of doing? Why didn't he even ask before going straight to Boss Han?
Zhao Wenyuan was puzzled and asked, "What is it?"
Jiang Jiusi said, "My friend is a researcher and has achieved some results, but she's recently hit a technical bottleneck and hasn't made a breakthrough in a long time. After hearing about her difficulties, I feel that you, President Zhao, might be able to offer some assistance."
"If it's a research collaboration, just tell me directly; I can make the decision," Zhao Wenyuan said. "Since it was President Jiang who introduced us, I believe there must be value in collaborating."
Hearing this, Jiang Jiusi showed a slightly embarrassed expression and said, "Cooperation is one aspect, but she's also having some financial difficulties. I was thinking, why not try to involve both parties and see if Origin Group could also invest some money? It's not that I don't want to invest myself, it's just that the amount of money in my company is limited. I can co-invest if I want, but if it's just me, even if I put it all in, it might not be enough."
Zhao Wenyuan then understood why Jiang Jiusi insisted on finding Boss Han.
Zhao Wenyuan couldn't make the final decision on the investment. But Origin Group wasn't a company that specialized in investments, so why did they come to them?