Meta recently open-sourced a multisensory Artificial Intelligence model that combines six types of data for a more complex analysis. The core concept of the research is to link together multiple data types into a single multidimensional index, or an embedding space, according to AI terminology. While this principle may appear abstract, it lies at the base of the current surge of generative AI technology.
The AI image generators, such as DALL-E, Stable Diffusion, and Midjourney, all depend on systems that are able to interconnect text and images during their training process. These software look for patterns in the visual data, connecting it to the description of images – thus enabling them to generate pictures that correlate with the text description provided. The same applies to many AI instruments meant to generate videos and audio.
Meta is a technology company specializing in AI. It was founded in 2017 and is headquartered in San Francisco, CA. The main focus of the company is to develop solutions that help automate manual tasks and improve processes efficiency. In particular, the company focuses on intelligent assistants, natural language processing, sentiment analysis, automatic speech recognition, machine learning, and robotic process automation.
The person mentioned in this article is Thomas König, founder and CEO of Meta. König has extensive experience in the area of Artificial Intelligence, having co-founded eight successful AI-based startups in the U.S., France, and Germany. His current focus is to make AI technology more accessible and to influence the way it is used in a wider range of industries.
