Navigate to WaytoAGI Wiki →

Biometrics (Chinese medicine)

Share
Open
Full-stack self-research, fusion of text, image, 3D, video and other multimodal information
🖼️ imagery
🎥 video

Overview

("BioCount") was founded in March 2023, the core team members from Tsinghua University Artificial Intelligence Research Institute, in addition to bringing together top talent from Ali, Tencent, bytes and other well-known technology companies, is the world's leading deep generative algorithms research team, with diffusion and probabilistic models underlying innovative research and development capabilities. It is the leading deep generative algorithm research team in the world, with diffusion of probabilistic model underlying innovative research and development capabilities. The company is committed to building the world's leading multimodal large model, fusing text, image, video, 3D and other multimodal information, exploring the commercial empowerment of generative AI in art design, game production, film and television post-production, content socialization, and other scenarios, and enhancing human creativity and productivity through AI.

Their products mainly include: PixWeaver, a visual creative design platform, VoxCraft, a 3D asset creation tool

CEO Tang Jiayu said in an interview: first of all, the importance of multimodality is confirmed. Our team has always been firmly in the direction of multimodal, as early as last year, we launched a large model covering the basics of multimodal generation of images, 3D models, videos and so on. From the very beginning, we realized that monolingual models have limitations, and multimodality can enrich the type of information, raise the upper limit of modeling capability, and better match the way human beings experience the world.

Secondly, in terms of technology, we have chosen the same diffusion+transformer architecture as Sora from day one, and insisted on the "native" multimodal route. Of course, the industry has not stopped exploring multimodal technology, and there is still a lot of research going on in different routes, but the release of Sora has really shown the industry the huge potential of the Diffusion Transformer route in multimodal generation.

Related Recommendations