When compared to other companies, it is undeniable that Apple adopted a more reserved and mysterious stance, especially when it comes to the technology sector and all the advances achieved in the area of artificial intelligence.
Recent publications made by researchers linked to the company indicate that Apple is committed to developing new features for its iPhones. Information was provided about a long-awaited chatbot and a project centered on digital avatars.
The latest document reveals that Apple is working on an innovative approach to operating Extended Language Models (MLE) on devices with limited DRAM memory capacity.
This would make it possible for artificial intelligence systems to operate directly on iPhones or iPads, instead of relying on external servers.
Image: Reproduction
Models are driven by AI
These language models, similar to those that power chatbots like ChatGPTare generally processed in data centers with computing capacity much greater than that of a mobile device.
Apple researchers have discovered an answer to a computational hurdle by developing a model that tunes flash memory performance.
This strategy aims to reduce the amount of data transferred, reading information in larger, sequential blocks.
Apple has experimented with this method with models like the Falcon 7B, a scaled-down version of an extensive open-source language model.
As indicated by the researchers, strategies of this type will be fundamental to realizing the full potential of these language models across a wide variety of devices and applications.
The recent launch of MLX, an Apple framework aimed at developing machine learninghas sparked speculation about the feasibility of creating an AI-based generative application for the future.
This application would make use of Apple hardware, thus reducing the need for external processing.
Apple chatbot?
In previous reports, it was suggested that Apple was developing its own chatbot, possibly based on the extensive language model known as “Ajax”.
Although it has been the subject of internal tests, there is still no defined strategy for its launch on the market.
Several companies, such as Google with its Gemini system and OpenAI with initiatives such as the “iPhone of artificial intelligence”, are exploring the development of mobile devices with artificial intelligence capabilities.
It is predicted that by 2024, more than 40% of new devices will offer functionalities of this type.
The second study released by Apple, entitled HUGS, reveals a new research project that encompasses an artificial intelligence generating tool capable of quickly converting a brief video of a person into a digital avatar.
This method makes it possible to create animated digital avatars from short videos, with possible applications in social networks, games and augmented reality.
Cristiano Amon, CEO of Qualcomm, anticipates an increase in mobile device sales in early 2024, driven by the integration of generative artificial intelligence.
Amon highlights the prospect of more advanced virtual assistants capable of anticipating users’ actions, and eagerly awaits progress in photo editing due to these technological advances.
Support our work ❤️
If you enjoyed this article, consider leaving a tip to help us keep publishing great content.
