Summary: Ai new program developed by computer scientists identifying unaware journals to identify books that are not sitting. These magazines often charge high money to publish without proper peers, scientific reliability.
AI has analyzed more than 15,000 magazines and is violently violated, giving researchers in a light-way to see the risk. While the system is not perfect, it works as the first important filter, and human professionals make last calls.
Key facts
- The Dixed Publication: Magazines exploit researchers about charging money without any peer quality review.
- AI: The program is slaughtered for more than 1,000 of 15,200.
- Scientific firewall: It helps the maintenance of the prospect of research protection.
Source: University of Colorado
The University of Computer Scientist led by the University of Colorado Boulder has established a new platform for an automated electricity that automatically seeks “unorgld” scientific journals.
Research, published Aug. 27 From the Book:The Science Issue,“Dealing with the awesome practice in the research world.
Daniel Acuña, the leader of the leading copy of a computer science, receives a reminder of the number of e-mail inbox, and offered to publish his or her paperwork.
Such books are sometimes called food “in journals”. They look at scientists, reassuring them that they pay bigger or thousands of dollars to publish their research without proper exposure.
“There was a growing effort between scientists and organizations to have magazines,” said Acuña. “But it is like whack-a-mole. You have one, then another, usually from the same company. They just form a new website and come up with a new name.”
The new AI tool automatically indicates the scientific journals, evaluate their websites and other online information on specific procedures: Are the editing board set by established researchers? Are their websites containing many language errors?
Acuña emphasizes that the tool is not complete. Finally, you think human professionals, not equipment, should make a last call from whether the book is not unconscious.
But when prominent figures may ask the legacy of science, to stop the spread of books more important than ever, he said.
“In Science, you don't start at the beginning. He built more than the study of others,” said Acuña. “So if the base of that tower is fall, everything is falling.”
To move down
When scientists include a new lesson in a respected book, the study is usually a habit called peers. External experts learn a lesson and check them in quality – or, at least, that goal.
The growing number of companies have wanted to avoid that process to convert. In 2009, Jeffrey Beall, LETA booklist, named the words “public books” to describe the books.
Usually, they intend to researchers outside the United States and Europe, such as China, India, and Iran – scientific institutions become small, and stressful and motivated by researchers to publish higher.
“They will say, 'If you pay $ 500 or $ 1,000, we will update your paper,'” said Acuña. “In fact, they do not offer any service. They just take PDF and send it to their website.”
A few different groups are looking for preventing the habit. Among them is a non-profit organization called open access directory (DOJ).
Since 2003, volunteers in DAJs crash thousands of books as banging based on the process of six. (Honorable books, for example, often include a detailed description of their peer-peer-to-peer policy.)
But being accompanied by the spread of those books is difficult for people.
Running the process, Acuña and his colleagues have turned to Ai. The team trained its program using DOJ data, and then asked AI to have been a list of approximately 15,200 open open books.
In those magazines, AE initially threw more than 1,400 as a problem.
Acuña and her colleagues ask for people's experts to review the set of suspicious books. AI has made mistakes, according to people, and the books of approximately 350 books as it is possible. That is more than a thousand magazines in advance researchers who have identified as impressive.
“I think this should be used as a helper to get big statistics of magazines,” he said. “But human professionals should make the last commentary.”
Firewall of Science
Acuña added that researchers did not want their system to be “a black box” as other platforms in Ai.
“With Chatgpt, for example, often you don't understand why it lifts something,” said Acuña. “We have tried to make our own as translating as much as possible.”
The group was found, for example, that the unusual textbooks published the unusual number of articles. They also added the authors with a large number of meetings, as well as the authors who receive their research, research for other scientists, at a higher level.
The new AI program is not available in public, but researchers hope to make it available to university and copyrights. Acuña sees a tool as one where researchers can protect their fields of bad information – what he calls “science firewall.”
“Like a computer scientist, I often give an example that a new smartphone comes out,” he said.
“We know that phone software will be mistaken, and we look forward to repairs to future mistakes. We should probably do the same science.”
In terms of this AI and Science lesson
The author: Daniel Shurain
Source: University of Colorado
Contact: Daniel Strooin – University of Colorado
Image: This picture is placed in neuroscience matters
Real Survey: Open access.
“It is estimated to predict open-open books” by Daniel Acuña et al. The Science breaks down
Abstract
It estimates predictable booklisters
Discussing journalists threatening worldwide research, but manual vetting can be slow and unavoidable.
Here, we test the artificial intelligence skills (AI) such corrections in such areas by analyzing the construction of website, content, and a punction metadata.
Assessed against the broader information of the person specified by the person, our approach has achieved practical accuracy and previous blocking of the references to the Journal Legitimitacy.
By changing the limit of the decision, our way may prioritize a complete or accurate test, low sound.
On a balanced road, it is very cold, printing hundreds of thousands of articles, receiving millions of quotes, allowing funds from bigger agencies.
Error Analysis Produces prohibited articles, a series of books as bad as a magazine, and public shops for the limited range of the Internet, which is the quality-related issues.
Our acquisition shows the ability of AI to be considered honest, while and highlighted the need to pair the drilling triage with the expert.
