What’s New in Google’s AutoML
Google is finally up with their beta version of AutoML , which is going to be the milestone in the field of deep learning as we are doing since ages.
Google’s very new AutoML is afresh developed cloud software in the area of Deep Machine Learning tools. It is basically working on the Google’s fluent state of art which is research in image recognition named Neural Architecture Search (NAS). Neural Architecture Search is providing an algorithm which will eventually give a dataset. AutoML then give suite to machine learning tool which will allow only easily trained high- end performance for deep networks, and in this there is no need to have any kind of knowledge in the area of deep learning or Artificial Intelligence for the user, the only requirement for these kind of performance is just a labelled data. Google will take help from NAS to will find out the result to get best network for the users specific dataset and task. As they have already given many examples that how their ways can help the user to achieve best performance which are much more better than hand- designed networks.
Deep learning and Artificial Intelligence are together is massively powerful technology that we should never try to supress it under any high paywalls. Yes you are correct Google, Apple, Microsoft, Facebook and Amazon need to make money so that they can survive and compete to the market. But in case if the research paper is public where we have deep learning libraries available to get replication done very quickly, it is not making any sense to try and block all the details from people where all the information are so easily available openly.
This is having potential which is very much bigger than problem: There is one thing which is very trending in the area of Artificial Intelligence in most recent times and also the topic of discussion in much of the research community so that they can quickly post their work on sites such as Arxiv which will share the information with the community and surely get the feedback. So there is also a the trend of posting the research codes on GitHub so that they can reproduce and can be used in general in further algorithms in research and reality based applications.
Science is contributing their best part in progress and developing great general knowledge for every individual.