DeepSeek Might Have Just Killed the Text Tokeniser
4 Articles
4 Articles
Instead of text tokens, the Chinese AI company Deepseek packs information more efficiently into images. This makes it as good at important benchmarks as the top models. read more on t3n.de
Artificial intelligence has a short-term memory – and that's what slows its progress. Instead of decomposing language into countless text tokens, the Chinese company Deepseek transforms its AI information into images. Using a novel Optical Character Recognition System (OCR) machines are supposed to learn to remember more efficiently and longer. The post The End of Forgetting – DeepSeek lets AI think in images first appeared on ingenieur.de - Job…
DeepSeek Might Have Just Killed the Text Tokeniser
Researchers at DeepSeek have introduced DeepSeek-OCR, a new model that explores how visual inputs can help large language models (LLMs) handle longer text efficiently. Instead of feeding text directly into a model, DeepSeek-OCR compresses it into visual tokens, which are essentially ‘images of text’ that carry the same information in fewer tokens. This approach, called contexts optical compression, could help LLMs overcome one of their biggest l…
DeepSeek May Have Found a New Way to Improve AI’s Ability to Remember
(MIT Technology Review) – An AI model released by Chinese AI company DeepSeek uses new techniques that could significantly improve AI’s ability to “remember. Released last week, the optical character recognition (OCR) model works by extracting text from an image and turning it into machine-readable words. This is the same technology that powers scanner apps, translation of text in photos, and many accessibility tools. (Read More)
Coverage Details
Bias Distribution
- There is no tracked Bias information for the sources covering this story.
Factuality
To view factuality data please Upgrade to Premium

