Google’s Machine Learning AI Can Create Machine-learning Code Better Than The Researchers Who Made I
Google’s AutoML system has recently produced a series of machine-learning codes with higher efficiency than those made by the researchers themselves.
The company’s AI project, AutoML, which debuted in May, was developed as a solution to the lack of top-notch talent in AI programming. There are very few experts with the knowledge to build highly complex AI systems. So, to meet the demand for experts, the Google team came up with a machine learning software that can create self-learning code, and in a way, clone itself. The system can run thousands of simulations to determine which areas of the code can be improved, to make appropriate changes, generate new architectures, give feedback, and continues the process ad infinitum, or until its goal is reached.
Now, AutoML has outdone human engineers by building machine-learning software that’s more efficient and powerful than the top human-designed systems.
Google reported on its official blog that AutoML scored an 82 percent efficiency while sorting images based on their content. The system also beat human-made software at a more complex task of marking the location of several objects in an image. The code that AutoML wrote managed to score 43 percent while the best human-built program scored 39 percent. It means, the machine programmed better AI software than even the Google researchers could design.
Yet the AutoML software can only write programming for relatively basic AI tasks at the moment.
The world of artificial intelligence (AI) is getting broader each day with many big companies like Amazon, Apple and Facebook are actively contributing to the nascent field with their efforts. Now, this achievement by Google marks the next big step for the AI industry. With this breakthrough project, Google is surely taking AI to the next level.