Meta Unveils AI Tool for AI Assessment Without Human Input

Meta announced on October 19th, 2024, that its research team has introduced a series of new AI models, including a tool named the “Self-Taught Evaluator.”

This tool aims to reduce reliance on human intervention in AI development processes.

Initially presented in an August research paper, it employs a “chain of thought” technology, similar to OpenAI’s o1 model, enabling AI to reliably judge its own output.

This technology enhances accuracy in solving complex problems by breaking them into logical steps, showing promising improvements in fields like science, programming, and mathematics.

Meta’s researchers trained the evaluator entirely with AI-generated data, eliminating human involvement from the process.

The use of AI to evaluate AI demonstrates the potential for autonomous AI agents that can learn from their own mistakes.

Two researchers leading the project indicated that many AI experts foresee the future development of highly intelligent digital assistants capable of handling numerous tasks without human interference.

These self-improving models are expected to reduce the current need for “reinforcement learning from human feedback” (RLHF), a process that is often expensive and inefficient due to its reliance on experts to label data and verify answers to complex problems.

“We hope that as AI evolves, it will surpass human capabilities, gaining the ability to check its work and ultimately achieving accuracy beyond that of an average human,”

said Jason Weston, one of the project’s researchers.

“The ability to self-train and evaluate is key to achieving superhuman AI.”

Reports indicate that other tech companies like Google and Anthropic are also working on the concept of “reinforcement learning from AI feedback” (RLAIF).

However, unlike Meta, these companies usually do not publicly release their research models.

Meta’s newly released tools also include updates to its image recognition model “Segment Anything,” a tool to accelerate the response time of large language models, and datasets to help discover new inorganic materials.

Keep visiting for more such awesome posts, internet tips, and lifestyle tips & remember we cover,
“Everything under the Sun!”

Follow Inspire2rise on Twitter. | Follow Inspire2rise on Facebook. | Follow Inspire2rise on YouTube

Aditya Nath Jha

Aditya Nath Jha is an Engineer from New Delhi, India. His areas of interest include Gadgets, WordPress, speed optimization & latest technology. When he is not busy blogging he loves to write poetry, compose his own songs and has a taste for music! Find him on Facebook, Twitter, Linked in, Instagram. And watch his videos on YouTube.