ChatGPT may have caught the world's attention. But DeepMind, the Google-owned research lab, claims its next big language model will rival — or even surpass — OpenAI's.
See also: DeepMind AI: Solved a decades-old mystery about proteins!
According to a Wired report, DeepMind is using techniques from AlphaGo, DeepMind's AI system that was the first to beat a professional human at the board game Go, to build a chatbot that competes with ChatGPT and is called Gemini.
If all goes according to plan, Gemini will have the ability to design or solve problems as well as analyze text, DeepMind CEO Demis Hassabis told Wired's Will Knight.
Knight speculates that Gemini, which was briefly shown at Google’s I/O developer conference in May, will leverage innovations in reinforcement learning to accomplish tasks that current language models struggle with. Reinforcement learning involves “rewarding” an AI system for certain behaviors and/or punishing undesirable behaviors, with the goal of “teaching” the system what behaviors it should exhibit in a given situation.
Proposal: DeepMind developed the AI AlphaTensor to perform matrix multiplication calculations
As Knight notes, reinforcement learning has already led to gains in language modeling—it’s key to how systems like ChatGPT respond to prompts. DeepMind, with its wealth of experience in reinforcement learning ( AlphaGo is one example), is no doubt eager to apply its knowledge to the field of genetic AI.
It’s worth noting that Gemini isn’t DeepMind’s first foray into language models. Last year, the company introduced Sparrow, a chatbot that the lab claimed was less likely than other language models to give “unsafe” or “inappropriate” answers to questions. Hassabis told Time in January that DeepMind would consider releasing Sparrow for a private beta sometime this year—it’s unclear if those plans are still in the works.
Gemini is DeepMind's most ambitious generative AI to date. The project was spurred by the failure of Bard, Google's chatbot project, to keep up with ChatGPT, and has direct involvement from senior Google executives. According to a study, the generative AI market could reach $109.37 billion by 2030.
Read also: DeepMind: We trained an AI to control a nuclear fusion reactor
information source: techcrunch.com
