AI researchers named the world’s best laboratories in this field
Artificial intelligence (AI) scientists do not like being asked to name the best development labs in the field, because this question may be difficult for them to answer..
Major U.S. tech companies such as Google, Facebook, Amazon, Apple and Microsoft have built dedicated AI labs over the past decade. There is also DeepMind, owned by Google’s parent company Alphabet, and OpenAI, of which he is the founding investor. Elon Musk.
«I hate this question», – told CNBC Mark Riedl (Mark Riedl), assistant professor at the Tech School of Interactive Computing, when asked to highlight favorites.
«In terms of reputation, there is a good argument to say that DeepMind, OpenAI and FAIR (Facebook AI) are among the top 3», – he said.
Top ARTIFICIAL INTELLIGENCE RESEARCH LABS IN THE WORLD
Artificial Intelligence Investor Nathan Benaich (Nathan Benaich), a partner at Air Street Capital, agreed with him. According to him, Google Brain and Microsoft could potentially be included in the top lists, while, in his opinion, Amazon and IBM are not in the same league with them when it comes to the results and impact of AI research..
Another AI expert, who wished to remain anonymous because he did not have permission from his company to speak publicly, told CNBC that DeepMind, OpenAI, and FAIR are likely the top three research labs focusing on AI in particular in terms of announced funding. while IBM issues more patents. He said that the unknown the question remains about Baidu and Tencent when asked about Chinese tech giants and their development in this space.
Alphabet gives DeepMind hundreds of millions of dollars a year to do their work, while Microsoft has invested $ 1 billion in OpenAI – on top of the $ 1 billion that was contributed. founding investors. FAIR funding is less clear because Facebook does not disclose details.
DeepMind is best known for its AI AlphaGo, which beat the world’s top players in the ancient Chinese board game Go. There’s even a Netflix documentary about AlphaGo’s victory over the legendary South Korean gamer in Go. – Lee Sedolem (Lee Sedol). The company is now trying to focus on using AI to solve humanity’s biggest scientific problems, and at the end of last year, it made a breakthrough in biology focusing on protein folding. (Protein folding).
OpenAI has developed AI-powered gaming software that is capable of beating human players in games like Dota II. However, more attention is paid to from the press the company received for its generator AI GPT-3 text and its fancy generator AI Dall-E images.
FAIR, meanwhile, doesn’t have the well-known strong AI like AlphaGo or GPT-3. But the Facebook lab team has published research papers in areas of interest to the social network, including computer vision, natural language processing, and AI-powered conversational speech..
One way to measure the impact of AI labs is to see how many scientific articles it publishes at two of the largest AI conferences: NeurIPS and ICML..
In 2020, Google had 178 articles accepted and published on NeurIPS, while Microsft had 95, DeepMind had 59, Facebook had 58, and IBM had 38. Amazon had less than 30.
During the same year, 114 articles were accepted and published from Google on ICML, while DeepMind – 51, Microsoft – 49, Facebook has 34, IBM has 19 and Amazon has 18.
AI has been hailed as a technology that could spark a new industrial revolution and dramatically change the world we live in. But, at least for now, the technology remains relatively nascent and «narrow» in its capabilities – an AI that can play chess at a superhuman level, for example, cannot make an omelet.
DeepMind, OpenAI, and FAIR are considered the top three labs in part because of «strong PR games», – said Riddle.
He believes Microsoft, which does most of its AI work through Microsoft Research, could legally be included in the top lab rankings.. «For some reason, they sometimes fly below the radars.», – said Riddle. «Salesforce, Amazon, IBM have really very good research done, but again, they don’t make a lot of noise.».
Neil Lawrence (Neil Lawrence), former director of machine learning at Amazon Cambridge, told CNBC that Amazon doesn’t have a large, centralized AI research lab because it’s more focused on bringing technology to customers.
«I would say they did it very successfully», – said Lawrence, who is now a professor of machine learning at the University of Cambridge. «But if you look at (academic) publications as a yardstick, they won’t rank.».
Lawrence said Microsoft Research is the research lab he personally admires the most, but «Amazon is indeed leading the way in AI deployment … despite not having a (large) research lab».
He added: «DeepMind, OpenAI, and FAIR have definitely dominated the headlines. But here I wonder how much of the research they publish could have traditionally been done in universities».