Apple’s Cutting-Edge LLM Image-Editing Tool Turbocharges GenAI Initiatives

Apple’s Cutting-Edge LLM Image-Editing Tool Turbocharges GenAI Initiatives

Apple is making a significant push ⁢into generative AI (genAI) with the ‍introduction of Keyframer, a powerful new tool that allows users to ‍animate static images using text prompts.

Keyframer, as described in a recently published Apple⁢ research paper, is an⁣ impressive tool that enables users to create animated illustrations ​from static 2D images ⁣by ​simply inputting an SVG image, providing text prompts, and letting⁣ Keyframer build CSS animation code to‍ animate the original ‌image, with⁣ the option⁢ to fine-tune the creation.

Apple’s AI efforts are rapidly expanding, with a series of AI-related releases and breakthroughs, indicating the‌ company’s ⁣commitment to staying ⁣at the forefront ⁤of AI technology.

Earlier this year, Apple introduced MLLM-Guided Image Editing (MGIE), an ‍AI tool for pixel-level image editing that‌ allows users to‌ perform image edits using text commands, potentially even ‌using ⁣Siri to execute those commands.

Apple’s recent acquisition of the iWork.ai domain has sparked speculation about its plans to integrate genAI into its devices, while its machine learning teams have published 16 research​ papers and discussions focused on LLM development, ⁢healthcare, speech, and ​more.

In late 2023, Apple unveiled significant⁤ AI technologies, including the ability ⁣to run Large ‍Language Models (LLM)-based AI efficiently on devices, ⁤the ML Explore machine learning framework‌ for Apple Silicon, and Ferret,‍ which‍ optimizes​ machine learning, as well‍ as a model that rapidly creates avatars from video.

Apple is also heavily investing in AI start-ups, with ​reports of 32 acquisitions in the last ‌year, signaling its commitment to leading the ⁣AI⁤ arms‍ race.

Analysts believe that Apple’s intentional focus on specific AI domains, such as machine image intelligence and AI, will yield ​significant results in the long term.

2024-02-15 01:00:04
Link from www.computerworld.com

Exit mobile version