New York:

Meta, owner of Facebook, announced on Friday the release of a batch of new AI models from its research division, including a “self-taught evaluator” that could pave the way for less human involvement in the development process of AI.

This release follows Meta’s introduction of the tool in an August article, which detailed how it relies on the same “chain of thought” technique used by OpenAI’s recently released o1 models to enable it to make reliable judgments on the models’ responses.

This technique involves breaking down complex problems into smaller logical steps and appears to improve the accuracy of answers to difficult problems in subjects like science, coding, and math.

Meta researchers used entirely AI-generated data to train the rater model, also eliminating human input at this stage.

The ability to use AI to reliably evaluate AI offers insight into a possible path toward creating autonomous AI agents that can learn from their own mistakes, two of the researchers told Reuters Meta at the origin of the project.

Many in the AI ​​field view these agents as digital assistants intelligent enough to perform a wide range of tasks without human intervention.

Self-improving models could eliminate the need for an often expensive and inefficient process used today called reinforcement learning from human feedback, which requires input from human annotators who must have specialized expertise to label the data accurately and verify answers to complex mathematical and written queries. are correct.

“We hope, as AI becomes more and more superhuman, that it will be better and better able to verify its work, so that it will actually be better than the average human,” Jason Weston said , one of the researchers.

“The idea of ​​being self-taught and able to self-assess is fundamentally crucial to achieving this kind of superhuman level of AI,” he said.

Other companies, including Google and Anthropic, have also published research on the concept of RLAIF, or Reinforcement Learning from AI Feedback. However, unlike Meta, these companies tend not to release their models for public use.

Other AI tools released by Meta on Friday included an update to the company’s Segment Anything image identification model, a tool that speeds up LLM response generation times, and datasets that can be used to facilitate the discovery of new inorganic materials.

(Except for the headline, this story has not been edited by NDTV staff and is published from a syndicated feed.)