General Adversarial Networks are helping artificial intelligence systems “think”. General Adversarial Networks are a form of artificial intelligence using a process similar to knockout genes in biology. In short the system puts two neural networks against each other to achieve a larger goal. After being trained, General Adversarial Networks have its “neurons” representing features closely mirroring how humans represent concepts like trees or doors, as well as the nuances between types of trees or doors.
Humanity is growing its reliance on artificial intelligence systems automating different parts of our lives. General Adversarial Networks is one particular line of research making artificial intelligence systems smarter.