Another shot between Google’s AI Brain Trust and more disagreements

Less than two years after Google fired two researchers who criticized the bias in artificial intelligence systems, the company fired a researcher who questioned his work on the capabilities of a specialized type of artificial intelligence used to make computer chips.

Researcher Satrajit Chatterjee led a team of scientists to challenge a well-known research paper that appeared last year in the scientific journal Nature, saying that computers could create certain parts of a computer chip faster and better than humans.

Dr. Chatterjee, 43, was fired in March, shortly after Google told its team it would not publish a paper denying some of its claims in Nature. Matter. Google confirmed in a written statement that Dr. Chatterjee had “stopped for a reason.”

Google declined to give details of Dr. Chatterjee’s dismissal, but offered full protection of his research and willingness to publish his assessment.

“We have thoroughly reviewed the original Nature Paper and support the reviewed results,” said Zubin Ghahraman, Vice President of Google Research, in a written statement. “We have also strictly reviewed the technical claims for further submission and it does not meet our standards for publication.”

The dismissal of Dr. Chatterjee was the latest example of disagreement in and around Google Brain, the AI ​​research group that is key to the company’s future. After spending billions of dollars hiring leading researchers and creating new types of computer automation, Google has struggled with numerous complaints about how it builds, uses, and reflects these technologies.

Tensions between Google’s artificial intelligence researchers reflect a much larger battle in the tech industry, with many questions being asked about new artificial intelligence technologies and the thorny social issues that are entangled in these technologies and the people who created them.

The latest controversy will also be followed by dismissal and a duel of claims between Google’s artificial intelligence researchers, which is a growing concern for a company that’s its future in artificial intelligence. Sundar Pichai, CEO of Google’s parent company, Alphabet, likened artificial intelligence to the advent of electricity or fire, calling it one of humanity’s most important endeavors.

Google Brain started as a side project more than a decade ago when a team of researchers created a system that learned to recognize cats in YouTube videos. Google executives were so obsessed with the prospect that machines could learn skills independently, they quickly expanded the lab and laid the groundwork for re-creating the company with this new artificial intelligence. The research group has become a symbol of the company’s greatest ambitions.

Prior to his dismissal, Dr. Gebru sought permission to publish a research paper on how AI-based language systems, including technologies developed by Google, could use the biased and hateful language they learn in books and websites. Dr. Gebrum said he was outraged by Google’s response to such complaints, including its refusal to publish the article.

A few months later, the company fired another team leader, Margaret Mitchell, who publicly condemned Google’s management of the situation with Dr. Gebrus. The company said Dr. Mitchell had violated his code of conduct.

A paper published last June in Nature disseminated a technology called reinforcement training that, according to the paper, could improve the design of computer chips. Technology has been hailed as an achievement of artificial intelligence and a major improvement on existing approaches to chip design. Google said it used this technique to create artificial intelligence to create its own chips.

Google has been working on the use of machine learning techniques in the design of chips for years, and it published a similar paper a year earlier. Around that time, Google asked Dr. Chatterjee, who has a Ph.D. from the University of California, Berkeley and worked as a research scientist at Intel, to see if the approach could be sold or licensed to a chip design company. – say the experts in the case.

But Dr. Chatterjee made a reservation in an internal email about some of the newspaper’s claims and questioned whether the technology was severely tested, three people said.

While the debate over this research continued, Google sent another paper to Nature. People said that for the submission, Google made some adjustments in an earlier paper and removed the names of two authors who worked closely with Dr. Chatterjee and also expressed concern about the newspaper’s main claims.

When the new paper was published, some Google researchers were surprised. They believed this did not follow the approval process of the publication, which Jeff Dean, the company’s senior vice president who oversees most of his artificial intelligence efforts, said was necessary after Dr. Gebru’s release, people said.

Google and one of the newspaper’s lead authors, Anna Gold, who co-authored it with fellow computer scientist Azalea Mirhossein, said changes to the previous paper did not require a full approval process. Google has allowed Dr. Chatterjee and several internal and external researchers to work on a paper that casts doubt on some of his claims.

The group submitted a controversial paper called Resolution Committee to approve publication. A few months later the paper was rejected.

Researchers working on the rejection paper said they wanted to escalate the issue for Mr. Pichai and Alphabet’s board of directors. They argued that Google’s decision not to publish the denial violated its own AI principles, including adhering to high standards of scientific excellence. Soon, Dr. Chatterjee was informed that he was no longer an employee.

Ms. Goldy said Dr. Chatterjee had asked them to manage their project in 2019 and that they had refused. When he later criticized this, he said he could not substantiate his complaints and ignored the evidence presented by them.

“Sat Chatterjee has been running a disinformation campaign against me and Azalea for more than two years,” Goldi said in a written statement.

He said the work was reviewed by Nature, one of the most prestigious scientific publications. He added that Google has used their methods to create new chips and that these chips are currently used in Google’s computer data centers.

Lori m. Burgess, Dr. Chatterjee’s lawyer, said it was disappointing that “some authors of Nature’s work are trying to close the scientific discussion with Dr. Chatterjee slander and attack simply because of scientific transparency.” Ms. Burgess also questioned the leadership of Dr. Dean, who was one of 20 co-authors of Nature’s paper.

“Jeff Dean’s actions to stop the release of all relevant experimental data, and not just data that supports his preferred hypothesis, should be of deep concern to both the scientific community and the general public who consume Google’s services and products,” Ms. Burgess said. .

Dr. Dean did not respond to a request for comment.

After the controversial paper was shared with scientists and other experts outside of Google, controversy spread in the global community of researchers specializing in chip design.

Chip maker Nvidia says it has used chip design methods that are similar to Google, but some experts are not sure what Google research means for the larger tech industry.

“If it works really well, it’s going to be a really great thing,” said Jens Lining, a professor at the University of Technology in Dresden, Germany, referring to the AI ​​technology described in Google’s paper. “But it is unclear whether it works.”

Leave a Comment

Your email address will not be published.