What layoffs at Meta and Twitter's artificial intelligence and machine learning divisions mean

As part of Twitter’s massive layoffs, the company’s ethical AI team, which worked to make Twitter’s algorithms more transparent and fair, was laid off as a whole ten days ago. The team, called ML Ethics, is led by Rumman Chowdhury, known for her leadership in applied algorithm ethics.

At the same time, Meta laid off 11,000 employees last week, accounting for 13% of the company’s total employees, including one owner A team of 50 people focusing on machine learning infrastructure research is named Probability. A researcher on the Probability team said that in the team, 19 people are responsible for Bayesian modeling, 9 people are responsible for ranking and recommendation, 5 people are responsible for machine learning efficiency, and 17 people are responsible for artificial intelligence in chip design and compilation. and some managers.

Experts say the layoffs at the two companies are significant because they mark a changing landscape for even the hottest industries of artificial intelligence and machine learning, as well as for Big Tech. Reflections on how companies and business teams approach their own responsible AI efforts.

The leader of Meta’s artificial intelligence and machine learning team was fired

Georgios Gousios, research director at software developer Endor Labs and associate professor at Delft University of Technology in the Netherlands, said that Meta’s Probability team was quite A tactical squad of elite troops.

Gousios worked on the Probability team from October 2020 to February 2022. He said that Facebook has many developers working in departments such as technology stack and business. The work done by the Probability team is related to daily software development and aims to invent and apply new tools/methods to make other teams more efficient in their daily work. .

This includes probabilistic programming (writing programs that employ distributions rather than single values ​​to represent variables), differentiable programming (making neural networks more efficient), and software engineering applications such as using machine learning, he explained Tools that help engineers write code faster with fewer errors, and debug inevitable problems faster.

He said, "The technical level of the Probability team members is very high, many of us have been engaged in academic research for many years. Many have decades of research experience at places like Microsoft Research or Bell Labs , more than 60% of people have a doctorate."

Many industry personnel in the fields of artificial intelligence and machine learning expressed surprise at Meta's layoffs, because the Probability t team had been highly valued before.

These are some of the best researchers in the field, just not as well-known as others, said Nantas Nardelli, a senior research scientist at CarbonRe, a developer of artificial intelligence for climate technologies.

They tend to make products that are less glamorous, but that may become the mainstay of machine learning products in 5 to 10 years, he said.

He explained that their Machine learning jobs are well suited to problems where low- or medium-level data volumes, high domain knowledge, and estimation uncertainty are important. "This expertise is often hard to come by, and there are now an increasing number of people who specialize in it," he said. few. "

Twitter's cuts to its ethical AI team offer lessons for businesses

Triveni Gandhi, who leads artificial intelligence at data science and machine learning platform Dataiku, said she was disappointed with Twitter's ethical AI team as a whole. She was not surprised to be laid off.

She said in an interview with industry media: "My gut reaction is that of course they are the first team to be fired because the current leadership of Twitter company is very committed to ethics, trust Have different views on security issues. ”

But as a responsible AI executive, she said she also began to think about what this news meant for her enterprise customers. “Will they also start to think about not needing these technologies or applications,” she said. ?”

However, she said the reaction from industry players to the layoffs at the two companies shows how important and respected ethical AI has become.

She said: “I think , other companies see this particular team being cut, and they think, 'We don't want to make the same mistake again. We don’t want to create a sense of distrust among consumers who adopt our artificial intelligence products. ”

She added that she saw a sense of determination in comments from customers brought about by news of layoffs in Twitter’s Ethical AI team. “They said, ‘We can do better and we will,’” she said. Enable responsible AI teams and start getting them to put this technology into practice, and let’s get started. ”

The above is the detailed content of What layoffs at Meta and Twitter's artificial intelligence and machine learning divisions mean. For more information, please follow other related articles on the PHP Chinese website!