Stanford’s Views on AI’s ‘Foundations’ Are Trying to Argue


Last month, Stanford researchers have announced that a new season of artificial intelligence had arrived, one built at the very top neural networks and the oceans. Adati a new research site at Stanford builds — and reads these “basics” of AI.

Opponents of the idea came out immediately – in addition to a meeting organized to recognize the establishment of the new venue. Some oppose the limited potential and sometimes useless practices of these species; others warn against focusing too much on one way to make machines smarter.

“I think the word ‘foundation’ is very wrong,” Jitendra Malik, UC Berkeley professor who studies AI, he told those present at the meeting in a video chat.

Malik conceded that one of the most popular genres known to Stanford researchers – major types of languages ​​that can answer questions or make statements from the past quickly – is very useful. But he said the evolutionary biology shows that language is based on other intellectual factors such as globalization.

“These models are really skyscrapers; they have no basis, “said Malik.” The language we have in these parables is not stable, there are some errors, there is no real understanding. “

According to a report by many Stanford researchers, it was “a new and innovative process,” and it called it “the beginning of the end.” Major forms of AI have made remarkable progress in AI in recent years, in areas such as robot awareness and language.

Major linguistic models also form the basis of large modern corporations such as Google and Facebook, which are used in places like search, marketing, and quantity. Creating and teaching large languages ​​can require the computer power of a valuable cloud compiler; meanwhile, it has hampered their development with the use of a few heel-cutting companies.

But big examples are hard, too. Language groups receive racist and derogatory remarks from much of what they are taught, and do not understand or know the truth or falsehood. Soon, the major language species will disappear spit on unpleasant language or false. There is no guarantee that these major brands will continue to advance mechanical technology.

The Stanford concept has divided researchers. “Calling them ‘the foundation of the foundation’ completely disrupts the story,” he says Subbarao Kambhampati, professor at Arizona State University. There is no clear path from these models to most types of AI, Kambhampati says.

Thomas Dietterich, professor at Oregon State University and former president of Partnership for the Advancement of Modern Wisdom, says he has “great respect” for the researchers behind the new Stanford site, and believes he is genuinely concerned about the complexities these species bring.

But Dietterich wonders if the concept of original forms does not provide the financial support needed to build and use them. “I was amazed that he gave these colors a beautiful name and made it a capital,” he says. “This is similar to planting a flag, which can be very profitable.”

Stanford has also encouraged the establishment of Global AI Cloud Creating industry-leading tools to make it possible for professionals performing AI research projects.

Emily M. Bender, a professor in the Department of Languages ​​at the University of Washington, says he is concerned that the concept of pricing is showing bias in investing in industry-based AI preferences.

Bender says it is very important to learn about the dangers of large types of AI. He encouraged co paper, published in March, reported on the problems of multilingualism and contributed to the departure of two Google searchers. But he says monitoring should come from a number of factors.

“There are all these other neighborhoods, the most important of which are just hunger for money,” he says. “Before we throw money into the cloud, I want to see the money working on other projects.”



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *