As the Polyvit project shows artificial intelligence is becoming more and more similar to humans. This is also demonstrated by a current research area of the AI team at Google. For the first time, the researchers have succeeded in enabling the AI to evaluate images and sound simultaneously. This means that it will be possible for the machine to see and hear at the same time.
The significance behind this development is considerable. Currently, if an AI is to recognize voice and face in a video, two separate machine learning models have to be activated. Not only does this create more work, but it also produces different results because the two systems have been trained differently. The team of researchers from Google, the Alan Turing Institute and the University of Cambridge have found a way to perform multiple tasks simultaneously in a recent research advance. According to a recent scientific paper, the Polyvit project can perform image and audio recognition at the same time, leading to new cutting-edge results.
If further tests with Polyvit also turn out positively, the system can, for example, also be run on terminals with considerably less memory than is currently the standard. Updates to the algorithm will also be easier to perform and the system will be able to achieve much better results in the near future. The researchers are currently feeding new data sets into the system to further test its functionality, but they are well on their way!
In a few years, such new AI systems will be able to create faster and more detailed forecasts, which will enable new and previously unimagined areas of application. Research projects like Polyvit are therefore central to the future of our society, our work and our everyday lives.