Machine learning without programming is now possible
More and more initiatives allow SMEs to use artificial intelligence without the need for programmers. Giants like Baidu and Google, as well as smaller companies like Lobe, are presenting their products.
Some people think that artificial intelligence and machine learning are only within reach for large corporations with the vast resources —both human and computerized— needed to train and feed machine learning models. But that’s changing. One of the basic premises of growing digitization is that it’s accessible to everyone. In this case, that also means small and medium sized enterprises (SMEs).
This is Baidu’s aim with its latest creation. EZDL is the name of a software development platform created by China’s colossal Internet company (one of the most popular search engines in the word, also an AI power). It’s designed for those who have no idea how to program. “EZDL is a service platform that lets users build personalized machine learning models through a drag and drop interface. It only takes four steps to train a deep learning model, built specifically for your business needs,” explained Yongkang Xie, technical director of the project. “Even if you know nothing about programming, you can build models on this platform without any barriers,” he adds.
One of the benefits of EZDL for SMEs is that its models work with limited data. A small company doesn’t usually have millions of data points to feed an effective model, but EZDL works with hundreds —or even dozens— of data points.
For now, this platform is capable of building three types of models: image classification, sound classification and object detection. Baidu offers examples of Chinese companies that have already tried EZDL for everything from finding defects in the manufacturing of computer keyboards to recognizing the sounds of different species of animals (useful for scientific work) and identifying errors in product placement on supermarket shelves. EZDL is free for basic use, but charges once the volume of operations exceeds a certain limit.
Deep learning in your browser
Another global Internet giant charges for its solution from the start. Google’s Cloud AutoML trains codeless machine learning models, with a simple interface (also based on drag and drop). It focuses on artificial sight, natural language and translation. Google also offers Teachable Machine, an even simpler tool designed for amateurs interested in experimenting with machine learning and understanding how it works. With nothing more than a camera (webcam or cell phone camera), Teachable Machine feeds a small neuronal network in the browser, without having to send the images to a server.
The U.S. startup Lobe, which now belongs to Microsoft, has a similar project focusing on artificial vision. With Lobe, even users who don’t know what TensorFlow is, and couldn’t write a line of code if their life depended on it can take their first steps in machine learning to test their ideas. Dragging a series of examples from the computer desktop, Lobe trains the model and allows users to insert it in the app they are developing, for example. “People have ideas and want to try out machine learning, but don’t know how to make the prototypes,” said Mike Matas, Lobe Co-founder, who explains that their target audience includes professionals from a wide range of fields like architecture or astronomy.
All these platforms and tools still have their limits, but are a sign that machine learning for amateurs —simplified and almost homemade AI— is getting closer and closer to our computers.