Apple researchers have built a massive model of multimedia artificial intelligence by submitting a research paper that enables them to achieve advanced capabilities to teach artificial systems languages and enable them to understand them and analyze images.
This is after CEO Tim Cook said: “Artificial intelligence features could arrive later this year,” according to what was stated on the Gadgets 360 platform.
The research team confirmed that work is underway on MM1, which is a multimedia model that contains 10 billion parameters and is called “Multimedia LLM (MLLM).” To understand texts and inputs on images, devices have been prepared to encode images, a vision language connector, architecture components, and data selections have been created. For an AI model to understand both text and image-based inputs.
The training indicated that the use of image captions, overlapping image text, and text data are important for the success of the technology (SOTA), according to the data released. Although the achievement is significant, the research paper is not sufficient to ensure the inclusion of a multimedia chatbot equipped with artificial intelligence. There is still a long way to go for the tech giant to reach an original basic model for artificial intelligence, but it has already taken a big step.
ظهرت في الأصل على yallaapk.com