Apple

Apple Machine Learning Research has introduced STARFlow-V, a novel generative model leveraging normalizing flows to achieve high-quality, end-to-end video generation, offering a compelling alternative to diffusion models.

Apple Machine Learning Research has unveiled a significant pseudo — annotation pipeline designed to address the scarcity of high-quality data for sign language AI.

Apple Machine Learning Research has published findings explaining how conditional diffusion models achieve compositional generalization, enabling them to generate novel combinations of objects and attributes unseen during training.

Apple Machine Learning Research has unveiled StereoFoley, a novel framework capable of generating semantically aligned, temporally synchronized, and spatially accurate stereo audio at 48 kHz directly from video input.

Apple Machine Learning Research has introduced LaDiR (Latent Diffusion Reasoner), a novel framework that integrates continuous latent representation with iterative refinement via latent diffusion models.

Apple has introduced a new App Store subscription option allowing developers to offer discounted monthly rates in exchange for a 12 — month user commitment.