6thWave: AI News Hub

Google’s X Labs have developed an AI that mimics the human brain’s learning process using 16,000 processors to create a billion neural connections. The system was tested with 10 million images from YouTube videos, without any predefined instructions. Impressively, the AI independently learned to recognize cats by creating a digital image based on common features. This experiment highlights the potential of neural networks to mimic human visual cortex functions. The AI’s ability to identify cats was tested against 20,000 random images, achieving a 70% better accuracy than previous methods. This achievement demonstrates the promise of machine learning to enhance various technologies, including translation algorithms and semantic understanding, making Google’s products more intuitive. The research will be presented at a conference in Scotland, showcasing a significant step towards more human-like AI.

Source.