[ad_1]
There is an old one a joke that scientists often say: Everything has already been discovered and published in the Russian press of the 1960s, we do not know. As trivial as it is, jokes can accurately describe the situation. The volume of information is large and growing rapidly: The number of scientific documents sent on arXiv (the largest and most popular server before) in 2021 is expected to 190,000-It’s just a summary of what scientists have done this year.
Apparently we do not know what we know, because no one can read all the books even in their few places (including, in addition to notes, PhD notes, lab notes, photographs, white papers, notes, and reports). Of course, it is possible that in this mountain of papers, the answers to many questions are hidden, the resources to be ignored or forgotten, and the connections are hidden.
Practical wisdom is one way to solve this problem. Algorithms are able to analyze style without human supervision in order to find a relationship between terms that contribute to discovery knowledge. But much can be accomplished if we stop writing scientific literature whose style and structure have not changed over the past hundred years.
Mobile messaging comes with a number of shortcomings, plus access to all content as well legal concerns. But most of all, AI is not understand ideas and relationships between them, and are affected by the prejudices that are maintained, such as the selection of papers they review. It is difficult for AI – and, even, for any unfamiliar readers – to understand scientific literature in any way because the use of different words varies from one instruction to the next and the same words can be used with different meanings in different areas. The increasing variety of research means that it is often difficult to interpret the topic correctly using mixed words to find all the relevant papers. Making connections with (re) recognizing similar ideas is difficult even for the brightest.
As a result, AI cannot be trusted and people will need to re-evaluate everything AI does after mining, a tedious task that hinders the use of AI. To solve this problem we need to make scientific not only paper on machines but also machines-logical, by (re) writing them in a specific programming language. In other words: Teach machine science in a language they understand.
Writing scientific knowledge in programming language will dry up, but it will last forever, because new ideas will be added directly to the science library that the machine understands. In addition, when machines are taught a lot about science, they will be able to help scientists establish their logical conclusions; mistakes, disunity, cheating, and repetition; and clearly the connection. AI is an understanding of the laws of nature is much more powerful than data-driven AI, so scientific machines can help in the future. Advanced technology machines can help rather than replace human scientists.
Mathematicians have already begun to translate. He teaches computer math by writing theorems and proofs in languages like Lean. Dependent and facilitator to verify and use a program in which one can initiate mathematical ideas as objects. Using known facts, Lean can determine if what he is saying is true or false, thus helping mathematicians validate evidence and identify areas where their ideas are complex enough. When Lean knows more, he can do more. The program of Xena’s work at Imperial College London intends to place all graduate mathematics courses in Lean. One day, reference helpers will be able to help mathematicians do research by looking at their hypotheses and analyzing the basic mathematical knowledge they have.
[ad_2]
Source link
