The generative adversarial network of the future is a very well-known technology with high popularity. Ian himself became a top player in the AI field because of this technology. His other well-known contribution is the book "Deep Learning," which is a bible-level textbook in this field.

Therefore, Ian has mentioned the cause and effect of this technology more than once in interviews.

So Meng Fanqi is very clear that Ian actually didn't come up with this idea until several months later in the historical timeline.

Meng Fanqi never considered the possibility of a collision of ideas. In his opinion, there is still at least half a year of time difference.

But he didn't expect that his early proposal of residual networks, which surpassed human-level image classification capabilities, would lead to such unexpected chain reactions.

"Oh my god, dear God.." Ian immediately squatted down and grabbed his head, looking very frustrated.

In fact, it is not uncommon for academic ideas to collide. Sometimes, if there is a time difference of one or two months, people usually don't mind and it may even become a temporary story.

But that is the situation where one party is still reviewing the paper and the other party submits it later.

The time difference in the review process is usually acceptable in the academic community.

Read this novel and other amazing translated novels from the original source at the "pawread dot com"

However, if the other party has already completed the experiment, written the paper, and even published it before one's own article is submitted.

Then no matter what you do, it's basically useless.

In fact, this stage is when the collision of academic ideas is discovered with the least loss. Ian has not yet invested energy in officially starting the project or writing the paper.

He is still in the early stage of discussing and communicating with his mentor.

But Ian understands the value of this idea very well. The reason why he is frustrated is because he regrets not thinking of this idea a few months earlier.

"How much collision is there? How different are our ideas?" Ian vented his emotions with a few big movements and immediately started thinking about remedies.

When academic ideas collide, everyone's cars are ultimately different brands.

There are advantages to both sides, and sometimes the side that publishes earlier can inspire the other side to further their research.

"It's a pity." Benjio quickly browsed through the paper while shaking his head and sighed, "The content we just discussed is completely a subset of the other party's true ideas."

At the same time, the printer in the corner of the office started making "clicking" sounds, and sheets of A4 paper slowly fell. Ian knew that his mentor was printing the paper where their ideas collided.

"Has it been officially published in a very formal paper format?" Ian's last bit of hope was shattered.

"Look at this LaTeX template, it should have been submitted to this year's CVPR. The other party's timing may be much earlier than we thought."

"Let me take a look first." Ian took a step forward and quickly glanced at the abstract and several formulas.

After only two minutes, he sat down on the sofa in the office very frustrated, "The other party is not just a guy who wants to publish a short paper to occupy the space. They came prepared."

There are many people in the academic world who, no matter what idea they have, will write a two or three-page paper and publish it on arxiv without any hesitation. This makes them appear to be the first to discover this matter, and then slowly conduct experiments.

This approach cannot be said to be wrong, but the atmosphere is really bad.

Ian has always believed that his sense of smell is quite cutting-edge. Even though he collided with the other party this time, it was just the idea of the other party wanting to occupy the space. He must still have a lot of room for maneuver.

But unfortunately, Meng Fanqi completely destroyed his fantasy with a very meticulous paper.

Not only were the definitions of the concepts and the derivation of the formulas very concise and clever, but the results based on DreamNet's related generative experiments were also unprecedentedly stunning.

"DreamNet, isn't that the method that won the top spot in IMAGENET2013 a few days ago?" Ian suddenly realized, feeling that he finally understood.

Ian was inspired by the breakthrough in human performance by this network, and he thought of using another network as a discriminator to enhance the performance of the generator through spiral confrontation.

Since he was inspired by this matter, the person who designed DreamNet, the backbone network, a few months earlier must have a deeper understanding than himself.

Then, following the trend, he conducted generative adversarial experiments, which was indeed reasonable and expected.

Thinking of this, Ian sighed deeply.

Perhaps if he lost to anyone else, he would feel very frustrated. Only losing to the author of DreamNet made him feel understandable.

After all, he was also inspired by the performance of this network.

"You have already achieved great success in max-out, Theano, and other things. How come you collapsed after a collision of ideas?" Benjio saw his student deflate like a balloon being pricked with a needle and couldn't help but find it a bit funny.

"We old guys have come a long way from the old days when AI was not respected." Benjio activated the universal skill of old people and reminisced about the past.

"Ten years ago, how many people were willing to acknowledge our neural networks? Don't be discouraged, cheer up."

Although he comforted him like this, deep down, Benjio also felt very heartbroken.His vision was far broader and deeper than that of Ian, who was currently in a state of excitement and elation. As he had said before, this technology had great potential for development and application in the fields of realistic face generation, real objects, painting, and even video-related areas.

Although the prototype was just a generator, it could be transformed into a style converter, a face modifier, with just a little modification.

It was a pity that just as he had the idea, someone else's paper had already been published with detailed formula reasoning, experimental results, and even application prospects in multiple fields.

"This might be a blessing in disguise. After all, nothing has started yet." Ian was depressed for a while, but still forced himself to cheer up.

He encouraged himself, "If I saw this paper when I was almost done, I really don't know how upset I would be."

Turning his head, he looked at the author column of the paper again. He really couldn't figure out where this Meng came from.

However, compared to the obscure HuangTang in the competition team, this DeqingFu seemed to be quite well-known.

Ian searched for Fu Deqing's information on his phone, and the more he looked, the more he frowned.

"Isn't this a professor in the mathematics department? Why did he come to mess around in the cutting-edge field of computer science!" Ian was quite indignant, and he even doubted whether he had checked the wrong information.

But after checking and checking, he couldn't find anyone else with the same name who seemed to fit the bill.

"Meng, I'll remember you! Next time, I'll definitely beat you to it." Ian had no choice but to accept reality and turn his grief and indignation into motivation.