AI & Enterprise
DeepSeek unveils Engram technology to boost AI inference efficiency, seen used in V4
China\'s AI startup DeepSeek published a paper on Engram, a new AI technique aimed at reducing computation by fetching simple factual information through conditional lookup from a separate memory store. The paper lists founder Liang Wenfeng, DeepSeek researchers and Peking University researchers as co-authors. DeepSeek suggested it may apply Engram to its next-generation V4 model and released implementation code on GitHub to strengthen open-source collaboration.
