Google’s TurboQuant AI-compression algorithm can reduce LLM memory usage by 6x

Google Research recently revealed TurboQuant, a compression algorithm that reduces the memory footprint of large language ...

MIT study suggests astrocytes play key role in brain memory storage

The human brain holds a staggering number of connections, yet scientists have long struggled to explain how it stores so much ...
Medical Xpress

Scientists take a deep look into memory processing inside the hippocampus

Resembling a seahorse, as its name implies from the Greek words "hippos" (horse) and "kampus" (sea monster), the hippocampus is a brain region crucial for memory formation. But until recently, ...
The Next Web

Google’s new compression algorithm cut memory stocks within hours of publication

Google's TurboQuant algorithm compresses LLM key-value caches to 3 bits with no accuracy loss. Memory stocks fell within ...
BioTechniques

Mitochondria of memory: the mitochondrial process supporting social recognition

This mitochondrial process supports the brain cells essential to learning, memory and social recognition. We know that mitochondria are the powerhouses of the cell, but what role do they play in ...

Nvidia says it can shrink LLM memory 20x without changing model weights

Nvidia's KV Cache Transform Coding (KVTC) compresses LLM key-value cache by 20x without model changes, cutting GPU memory ...

Just 20 minutes of physical activity may benefit your memory

A new study suggest that brief exercise, such as 20 minutes of moderate cycling, can boost brain activity, which may help to ...