Google Teachable Machine 2.0 Makes Learning AI Easier Than Ever
With Teachable Machine, anyone can create machine learning models fast and easy — no coding required.
It’s time we acknowledge that we’re all living in the age of Artificial Intelligence (AI). Dominating almost all the industries and sectors now, AI has drastically changed our lifestyle. It inevitably led to this rapid transformation in our way of living, and now, as a matter of course, learning Artificial Intelligence has never been easier — thanks to Google’s Teachable Machine 2.0.
Before this, people were always talking about the latest technologies and trends in machine learning and artificial intelligence. But few of them actually knew how it actually works. Now that the tech giant has introduced the site called “Teachable Machine 2.0”, we can get a comprehensive overview of what AI can and cannot do.
The best thing is, it does not require coding at all.
It means you can learn Artificial Intelligence in a fast and easy way and create interesting machine learning models without having to understand coding or programming.
Google has simplified Artificial Intelligence to such an extent where we can now train our computers using just a couple of clicks to make it recognize our audio, images, and poses. Yes, Google has made it that simple and easy to grasp the machine learning basics! Not only that, if you want you can also export your model to your apps or websites and make it come to life! Or else, just go ahead and save it on your own device for native use later!
How Does It Work?
With just a webcam and a microphone, Teachable Machine 2.0 can recognize your photos, poses, sounds, gestures, and patterns in minutes. For example, it can distinguish if you’re sitting down or standing up.
Google’s own open-source machine learning framework, the Tensorflow.js library lets you develop your machine learning models that can directly run in your browser or in your JavaScript run-time environment.
Build better voice apps. Get more articles & interviews from voice technology experts at voicetechpodcast.com
It means Google has made it super easy for us to train our machine learning models that work wherever JavaScript runs. Therefore, we can now create these models and use them in our own projects! We now have plenty of tools as options including Node.js, Glitch, P5.js and many more!
The interface of Teachable Machine 2.0 is as simple as it can get. There are only two buttons — one to record your data and another one to train your model. Just a click on the button launches the model training in your browser and the performance metrics are shown in the subsequent result panel. Once you’re done, it gives you options to export your achievements to your websites or apps or to third-party sites, or Google Drive or to your device. It lets you train and define more than three classes using files from your computer or from your webcam or mic.
Teachable Machine 2.0 Privacy Implications
For people just starting to work with machine learning, it can feel a bit intimidating and they get nervous that it may be over their heads. Google’s Teachable Machine 2.0 takes away that fear and sets the pathway for the new Machine Learning (ML) practitioners. They can quickly create prototypes on Artificial Intelligence without dealing with codes.
However, it’s common for people to worry about privacy. Google gives us reassurance regarding those privacy aspects too — saying it takes care of user’s privacy by running the entire thing on the user’s local device. It implies that no data you upload onto Teachable Machine 2.0 ever leaves your device. All your training samples stay on your device until and unless you choose to save it onto Google Drive.
Experiments and Applications
People have already started building interesting AI projects with the help of Teachable Machine 2.0. Steve Sailing, a technology expert has developed a communication tool, specifically designed for people with speech imparity.
Another one is the Teachable Snake Game, inspired by Webcam Pacman, built using React.js. Blakeley Payne, an Educational Researcher has convinced her team to use Teachable Machine 2.0 as part of middle-schoolers’ curriculum to enable hands-on experience on Artificial Intelligence.
Yining Shi, a Creative Technologist, and her students demonstrated the use of Teachable Machine 2.0 in the Interactive Telecommunications Program (ITP) at New York University to explore its potential for game design.
There are many other different kinds of projects built with Teachable Machine 2.0 such as Interplay Mode, Tiny Sorter and many more! There have been more than 1540 experiments with this and still counting. Overall, Teachable Machine 2.0 is a fun, interesting and functional platform for developing actual AI projects without having to code or write programs.
Bottom Line
Google’s Teachable Machine is more like an educational tool at heart. Hence, it may not be the most sophisticated Artificial Intelligence system at present. Nevertheless, the numerous experiments and projects we can create through this platform definitely make it a powerful tool for demonstrating AI in the real world, especially for the non-tech-savvy people.
To conclude, we can say that Teachable Machine 2.0 is not just an educational tool for Artificial Intelligence, but it’s a glimpse into our future of automation-driven lifestyle where people will be guided by devices loaded not just with artificial intelligence, but with “customizable” AI.
