What if Apple researchers had found a way to perfectSirito make it much more useful and relevant? Inan articlecalled ReALM: Reference Resolution As Language Modeling, scientists from the Cupertino company present a new tool that can take into account what is displayed on the screen and other tasks in progress.
Towards an AI revolution?
The icing on the cake is that this AI would be able to accomplish tasks very quickly and more powerful than OpenAI's ChatGPT 3.5 and ChatGPT 4.0, the benchmarks on the market at the moment.
In detail, the researchers of the apple brand explain as follows:
We demonstrate significant improvements over an existing system with similar functionality for different reference types, with our smallest model achieving absolute gains of over 5% for on-screen references. We also benchmark against GPT-3.5 and GPT-4, with our smallest model achieving comparable performance to GPT-4, and our largest models significantly outperforming them.
They add:“It also outperforms GPT-4 for domain-specific user utterances, making ReaLM an ideal choice for a practical reference resolution system that can exist on a device without compromising performance.”.
Of course, there is no question of catching fire yet and we will judge the effectiveness of this new technology. This presentation remains no less promising for the nextWWDC 2024which will start with the traditional keynote on June 10. During this event, which you will be able to follow with us, it is a safe bet that Apple will reveal more about its next innovations in AI.
i-nfo.fr - Official iPhon.fr app
By : Keleops AG
