A new machine learning solution from Google, the Cloud AutoML, allows businesses to develop custom models using the technologies and methods developed by the search giant. Cloud AutoML was developed after the tech firm observed that only a small number of businesses have the financial and human resources to take advantage of machine learning and artificial intelligence. By using the existing technologies created by Google, small businesses can now create models without the cost associated with constructing their own models. In addition, the search giant claims that its new platform could also help boost the productivity of experienced AI developers.
The AutoML Vision is the first product that the search giant has released as part of the Cloud AutoML platform. This product enables businesses to develop image recognition models for their own use case. Google claims that its AutoML Vision product offers increased accuracy due to the use of the company’s existing image recognition methods. Moreover, Google claims that its product can create a working AI model within a few minutes, although more complex models are usually completed in a day or longer. In order to produce an image recognition model, businesses should first upload at least a few dozen photographs. These photographs will then be used by the cloud to develop and train the image recognition model. For businesses who will upload unlabeled images, Google employs human labelers that will review the instructions provided by the business. The labelers will then categorize the uploaded images based on the businesses’ instructions. After the model is trained, users can access the predictions made by the model through the AutoML’s user interface or through a Rest API call. The Cloud AutoML product also offers integration with other services and products offered by the Google Cloud Platform.
Google has been very active in the development of artificial intelligence and machine learning solutions for developers and businesses within the last few years. Last year, the Mountain View-based search giant launched the Google Cloud Machine Learning Engine, a service that helps developers create models for any type of data. Also last year, the search giant launched the TensorFlow Lite, a machine learning solution that allows models that were developed and trained in the cloud to be incorporated into Android applications. The search giant promises that more services that cater to other major fields of artificial intelligence will be released in the near future.