Apple is no stranger to innovation, and their latest strides in artificial intelligence (AI) research are no exception. Recently, the tech giant has unveiled two game-changing techniques that could revolutionise the way we use our iPhones and iPads.
In the first research paper, Apple scientists introduce HUGS (Human Gaussian Splats), a method designed to generate animated 3D avatars from short monocular videos. This technique uses 3D Gaussian splatting to capture details like clothing and hair, and a novel neural deformation module to animate the Gaussians realistically. Compared to previous avatar generation methods, HUGS is up to 100 times faster in training and rendering, and it outperforms state-of-the-art techniques on 3D reconstruction quality.
The second paper tackles a key challenge in deploying large language models (LLMs) on devices with limited memory. Apple researchers have proposed a system that minimises data transfer from flash storage into scarce DRAM during inference. This breakthrough could soon allow complex AI assistants and chatbots to run smoothly on iPhones, iPads, and other mobile devices.
These innovations demonstrate Apple's growing leadership in AI research and applications. However, as they potentially integrate these technologies into their product lineup, they must consider the societal impact, from privacy protection to mitigating misuse. Nonetheless, these advancements could take artificial intelligence to the next level, making photorealistic digital avatars and powerful AI assistants on portable devices a reality.
Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai
Comments