Teachable Machine is machine learning without coding

Users can now have more fun with the upgraded Teachable Machine.
12 November 2019

Teachable Machine in action. Source: Google

Two years ago, Google launched an online site Teachable Machine, with the aim of educating the public on machine learning (ML) and artificial intelligence (AI). 

Users can train a simple model with the use of images without any coding required. However, there are some limitations to the model, which only allows users to train up to three classes through the camera.

The latest release lets users have a go at the full experience of training a model and let the projects come to life with real-life applications. 

Besides images, the new model lets users upload audio data and video clips as datasets for training. In their launch video, users are seen to upload files or pose in front of cameras live as inputs. 

Afterward, users can decide to export the trained model to embed it into other projects or save it on their devices for future use. 

Since the entire process, from providing input to the execution of projects, happens in one’s computer and users have full control of the datasets (they don’t have to leave the computer) the flexibility and freedom of the open-sourced ML platform has encouraged many to embark on projects solving problems in everyday life.

Projects from Teachable Machine 2.0

Various projects are emerging with the launch of Teachable Machine; one of them tells the story of Steve Sailing, who has amyotrophic lateral sclerosis (ALS), and experiments with the ML platform to develop a communication tool for better navigation and control of home appliances. 

Besides that, educationists are teaching the new generation the concepts of ML with Teachable Machine. For example, a middle school teacher, Blakeley Payne, uses Teachable Machine to educate students about AI with real-life demonstrations. 

Another, Yining Shi, incorporated the platform as part of the lessons and run school projects. Yining and her students built a gaming model using hand gestures to direct the avatar (see video below). 

These are some examples of the application of Teachable Machine, but ultimately the tool serves as a front door to understanding the endless possibilities the technology brings to communities and society as a whole.

Smaller businesses could use the tool as a way to understanding the power of AI and ML. Employees can be exposed to the breadth of the opportunity provided by the technology, which could help lay the way for larger investments in AI and ML in the future.  

For those a little more advanced, the model is powered by Tensorflow, Google’s open-source machine learning library, meaning it can be easily plugged into apps and websites. Organizations could use the tools to create simple machine learning applications, such as an AI assistant for internal use, or to create a data sorting machine, for example.