New KV cache compaction technique cuts LLM memory 50x without accuracy loss

MIT researchers developed Attention Matching, a KV cache compaction technique that compresses LLM memory by 50x in seconds — ...

Nvidia shrinks 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 costs and time-to-first-token by up to 8x for multi-turn AI applications.
BGR

Scientifically Proven Mnemonic Training Method Gives Your Brain Super Memory

Anyone can improve their memory using a proven scientific method the ancient Greeks and Romans developed. The idea is to create a "memory palace." A group of researchers says training daily using an ...
UPI

Ancient Aboriginal memory technique outperforms famous Greek method

May 18 (UPI) --New research, detailed Tuesday in the journal PLOS One, suggests an ancient memorization technique used in Aboriginal culture is superior to the "memory palace" method, an ancient Greek ...