Image Source: Unsplash
With the major intrusion of technological singularity in our current world, the potential of every AI we create, to perform specific tasks is beginning to explode.
Getting to the depth of technological singularity is resulting in developing a technology where machines will replicate their own self.
Google in its annual I/O developer conference, disclosed its one of the outlandish yet expected project, named as — AutoML. This is designed to solve the toughest part of creating an entire AI itself.
AutoML by Google and Facebook
It majorly revolves around fabricating the deep learning software. Which in precise is informed to choose appropriate architecture for a neural network.
Google has created a machine learning system which mainly utilizes reinforcement learning, which is supposedly a tried and tested technique. This is said to be the most common technique practiced by Google in its msot escapade.
This method will help researchers study and develop an efficient image processing and language recognition system.
When interviewed by MIT Tech Review, Google mentioned this project is long days away from becoming practically available.
Also, it pinpoints the definite usage of this technology that it will help look after or work up the tasks performed by human-machine learning and data science talents since they are slowly quitting from adopting the latest technologies.
Besides Google’s AutoML, Facebook is working on another super intelligent AI which is called, Automated Machine Learning Engineer as per WIRED.
It’s also called AutoML which identifies, chooses, and uses parameters that are proven to solve defects in a short span. Last summer, the AutoML Challenge became a huge hit as the game designers went on head-to-head to assemble machine learning ‘black boxes’.
As mildly expected, the machines successfully chose suitable models and tuned the parameters without any human mediation.
Game: Space Engineers
One of the winners — the crew behind Space Engineers have invested their profit into forming a team of experts who are creating an AI that will optimize its own hardware and software.
Though inventions like this will be of great help for non-experts to redistributeAI systems, which is good but is also expanding greater possibilities for machines to decide their own fate.
This is perhaps understood as ‘recursive self-improvement’ where the AI learns and transforms itself from a smart machine to super intelligent machine.
With the power of self-learning and self-improvising capabilities, machines will see massive growth in their intelligence in the proximal future.
Generally, the AI which is persisting in our current world, which is capable of doing what a human could, is only the ‘seed AI’ version.
Which is the basic level of AI and ML and that the creation of other super intelligent AI will certainly be the one with self-improving capabilities.