More than hardware and services, Apple has focused on developing increasingly specific tools that can manage huge amounts of data. Want an example? O Overton, one framework It is used internally by Ma to "automate the life cycles of artificial intelligence systems."
Ma engineers told the VentureBeat that the framework supports a variety of applications, helping them to process “track records” and answer “billions” of queries in multiple languages. According to these researchers, Overton's goal is to get developers to focus on the most advanced tasks and leave the data analysis schedule to the tool.
The focus is to help developers focus on higher-level tasks rather than lower-level machine learning tasks. These people can create learning-based applications without writing any codes. Overton (can) automate many of the traditional programming options, including the deep learning architecture, and allows the developer to build and monitor their application by helping with data manipulation.
More precisely Overton performs in a matter of seconds the processing of data (used in AI tools such as Crab), joined them to the model tasks, which are nothing more than the actions that the system must perform with that data. Despite being an internal tool, the framework Apple compiles information from many AI frameworks, such as Google's TensorFlow, Facebook's PyTorch, and Apple's own CoreML.
The tool also imbued with some UI related features such as the Slicing In addition to multitasking learning to predict which model is determined to use. So far, Overton has been very valuable to Apple researchers who have managed to reduce AI system errors by up to 3x.
In addition to Siri, Apple has also enhanced other features offered by its software by combining artificial intelligence and machine learning, such as the new iOS 13 detector that can identify dogs and cats in photos, among others.
via Apple World Today