Artificial intelligence systems might be reaching the state of “singularity” faster than expected. The state of singularity, in the context of AI systems, means that artificially intelligent machines will be able to create other artificially intelligent machines on their own without human involvement, and not just create artificially intelligent machines, but also create machines that are more intelligent than their predecessors.
Google’s “DeepMind” AI platform is one of the systems that has already begun to learn on its own, without human input. According to this NextWeb report DeepMind can now teach itself based on the information it already has. The Google system, called Differential Neural Computer, uses the existing data storage capacity to connect with other AI systems and use their existing knowledge to evolve further.
This video explains what Google is DeepMind system is:
The AI systems that simulate the neural networks of our brain use an interconnected series of nodes to stimulate specific knowledge centers in order to complete a task. They themselves can decide which notes to use in order to optimize a particular task execution or solution implementation. In order to deliver the desired outcome they can decide on their own, what method to use.
The DeepMind AI system uses experience to arrive at solutions. It uses its internal memory rather than outside conditioning and programming. For example, when the DeepMind AI system was given the basics of the London Underground public transportation system it was immediately, on its own, able to find additional routes and draw relationships between different routes.
It shows that the stage where artificial intelligence will be able to become more intelligent on its own, without depending on humans, is at our doorstep sooner than expected.