Shulou(Shulou.com)11/24 Report--

Tao Zhe-Xuan, a mathematical genius, also used AI tool GPT-4 in his research. With GPT-4, all you want to do is copy, paste and do it with one click!

Deep-Fried Chicken ChatGPT has also become a research tool for mathematical genius Tao Zhe Xuan.

Recently, he said online that he had found some small use cases of ChatGPT.

First, it is very good at parsing code-formatted documents (in this case, # arXiv search API), and then returns a properly formatted code query (later it also provides some working python code to call this API in the way I want, although I have to install a package manually to make it run). Second, I asked it to come up with some questions that smart students might ask in the undergraduate linear algebra class (for which I provided some sample questions), and it gave me some good examples of the possible direction of the course. and potential homework problems.

All in all, I find that AI tools such as ChatGPT can't help me solve an unsolved math problem, but they are useful for other work-related problems, although manual adjustments are often needed afterwards.

There is little difference between GPT-4 and GPT-3.5. How is GPT-4 good at parsing code-formatted documents?

Tao Zhe Xuan asked GPT-4 to search all arXiv articles whose metadata is a given keyword (for example, "Collatz") and submitted within a certain range of dates (for example, between 2011 and 2016). According to the following document, what is the format of such a query?

In response, GPT-4 not only gives the code, but also lists the categories of queries.

Netizens asked, as a reading tool, give a paper and ask it what questions the reviewer or speaker might ask? How is GPT-4 doing in this respect?

In this regard, Tao Zhe-Xuan took it to evaluate it and praised it as "really good."

"here are the first few pages of a recent paper. Can you ask some questions that the audience may ask when listening to the lecture on this paper? "

ChatGPT replied, "of course, here are some questions that mathematicians might ask after listening to a lecture on this paper." Then Barabara listed seven questions.

Careful netizens found that the screenshot of Tao Zhe Xuan's post was GPT-4.

He asked, "what do you think is the difference between the old ChatGPT and the new assistant based on the GPT-4 model? "

"so far, GPT-4 's reply is a little more useful and meets my requirements, but I don't see a significant difference in quality," Tao said. "

With the birth of ChatGPT and Workflow ChatGPT, many people are exclaiming that their jobs will be replaced by AI.

Even the talented mathematician Tao Zhe-Xuan has become familiar with ChatGPT.

On March 5, Tao Zhe Xuan said online that he had decided to try to incorporate AI tools into my workflow in different ways. Including GPT-4, DeepL and other tools.

During this period, he discovered the hidden functions of ChatGPT, such as finding formulas, parsing documents in code format, rewriting thesis sentences, and so on.

For example, ChatGPT can sometimes do a semi-finished semantic search in mathematics, that is, it can be used to generate hints.

For example, Tao Zhe-Xuan asked ChatGPT to recognize Kumer's theorem from the description, which did not give the correct answer, but according to its approximate answer (Legendre formula).

In this regard, Tao Zhe Xuan said that the role of artificial intelligence in mathematics is to provide a preliminary approximate answer, and then can be combined with traditional search engines to easily find the correct answer.

Tao Zhe Xuan also discovered the highlight of ChatGPT in dealing with mathematical problems, and was able to recognize transliterated versions of mathematical concepts in different languages.

In addition, you can ask ChatGPT to transform a bunch of references obtained from MathSciNet and format them as\ bibitems in the LaTeX bibliographic environment.

But doesn't ChatGPT ever make a mistake?

In a proof question about "whether there are infinitely many primes", Tao Zhe-Xuan found that the answer given by ChatGPT is not entirely correct.

On the other hand, he found that the idea of argument given by ChatGPT could be fixed, and he had never seen it before.

Copy, paste, one click to finish! Netizens also spent a lot of time mining the use cases of GPT-4.

Not only is it reasonable for GPT-4 to review papers, but it also hits the point.

As a scholar, I don't know what to think about this: I put one of my old papers in GPT-4 (divided into two parts) and asked an economic sociologist to conduct a rigorous but fair peer review. "

It created a perfectly reasonable peer review and stepped on many of the points raised by my reviewers.

For many people, if you want to do something, just ask GPT-4, "copy, paste, and do it with one click!" "

It's like using GPT-4 to make a complicated game.

Netizens formed a team of artificial intelligence assistants from GPT-4, Replit, MidJourney and Claude to create a 3D space parkour game from scratch.

However, Goose knows nothing about Javascript or 3D game programming.

Reference:

Https://mathstodon.xyz/@tao

This article comes from the official account of Wechat: Xin Zhiyuan (ID:AI_era)

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.