Google has broken down the way it uses AI and ML to deliver the best traffic data for its users in Google Maps. In a blog post, it mentioned the partnership with DeepMind, an AI lab of Alphabet to make predictions of the estimated time of arrival and journey.
Google says to be considering a lot of data including the history of traffic in such routes, suggestions from other drivers, etc to predict things in the best way possible.
Google uses AI and ML to Provide Accurate ETA in Google Maps
Over the years, Google Maps has grown to become so reliable service that, it’s now used by millions to travel over a billion kilometers every day. A major part that made Maps hit is the traffic data and routing suggestions, which Google has yesterday revealed in its blog post.
Google says it takes in a considerable amount of data from its users, history, and feeds them to its AI and ML models to get accurate predictions.
The predictions are so sophisticated that, they’re capable of estimating a certain route’s traffic in the near future now itself. Google says it takes the aggregate location data of its users (using Google Maps in live) to check how congested a route is.
While this gives an idea of how the traffic could be now, users need estimates in near future traffic, throughout their journey. For this, Google uses historical traffic data, which is fed into Machine Learning to predict future outcomes. Further, it said about collaboration with DeepMind, an Alphabet AI research lab to improve the traffic predictions.
Google says to be using a Machine Learning model named Graphic Neural Networks to make improved traffic predictions. This data is initially available to cities of Berlin, Jakarta, São Paulo, Sydney, Tokyo, and Washington D.C, and predicts a slowdown route even before you start. It also considers suggestions and marks from drivers, government datasets, etc to suggest better routes.
