Login

News Post

Researchers upend AI status quo by eliminating matrix multiplication in LLMs

ArstechnicaArstechnica · 4m
Researchers upend AI status quo by eliminating matrix multiplication in LLMs

Researchers have proposed a new method to run AI language models more efficiently by eliminating matrix multiplication, which is a core operation in neural networks. The study suggests that their approach could significantly reduce the environmental impact and operational costs of AI systems. The researchers developed a custom 2.7 billion parameter model without using matrix multiplication, achieving similar performance to conventional large language models. They also demonstrated running a 1.3 billion parameter model on a GPU accelerated by a custom-programmed FPGA chip, which consumed around 13 watts of power. The findings challenge the prevailing belief that matrix multiplication is essential for high-performing language models and could make large models more accessible and sustainable.

4
View Source

Comments

More Like this

Matrix multiplication breakthrough could lead to faster, more efficient AI models

Matrix multiplication breakthrough could lead to faster, more efficient AI models

A new trick uses AI to jailbreak AI models—Including GPT-4

A new trick uses AI to jailbreak AI models—Including GPT-4

Researchers say generative AI isn't replacing devs any time soon

Researchers say generative AI isn't replacing devs any time soon

How Big Tech sabotages rival startups — from the inside

How Big Tech sabotages rival startups — from the inside

Lessons from The PayPal Mafia: The Power of Being Open-Minded and Contrarian

Lessons from The PayPal Mafia: The Power of Being Open-Minded and Contrarian

From Phool to ZunRoof: 5 unconventional startups making waves in India

From Phool to ZunRoof: 5 unconventional startups making waves in India

Large Language Models’ Emergent Abilities Are a Mirage

Large Language Models’ Emergent Abilities Are a Mirage

Large language models can help home robots recover from errors without human help

Large language models can help home robots recover from errors without human help

LLM deployment platform Giga ML raises $3.6 million from Nexus Venture Partners, others

LLM deployment platform Giga ML raises $3.6 million from Nexus Venture Partners, others

NCLAT adjourns Byju’s case hearing to November 6

NCLAT adjourns Byju’s case hearing to November 6

Download the medial app to read full posts, comements and news.