is now everywhere, like an inescapable fog.
Why did this happen? Who knows? Maybe it was the horrific week of events that preceded its US launch, leading to some sort of mass psychological phenomenon of escapism. Maybe it’s the way viral trends seem to explode all the time, lately. Maybe it combines family-friendly, social and free-to-play ideas in some magic tonic.
Regardless, the moment has happened, now Nintendo is a company reborn. So what next?
I’m just kidding 🙂
I’m presenting the Watson GO game, It has been developed within the Watson IoT IBM Rome-Lab initiatives. The concept for the game was conceived in July 2016 by myself, just few days after Pokemon GO release date.
The development of Watson GO, in common with myself, is being led by the team:
Watson GO is a free-to-play image and location based game made in Bluemix, currently in alpha version.
To play you just need a Twitter account, here are the simple rules:
- every 4 hours Watson GO throws down a challenge posting the icon of an object present in the real world, any challenge has an #hashtag and a location
- you have to find the closest match to the target object
- walk around looking for a similar object (while you playing Pokemon GO)
- take a photo with you smartphone
- compose a tweet to Watson GO account with the #hashtag, the photo and your location
- see your score in the leader-board
- The goal is to find an object as similar as possible to the one posted by Watson
- Watson GO game uses IBM Watson image recognition APIs to find the closest match to the target object, plus score related your location
At the end of alpha testing, be the first of our leader-board, you will be rewarded with iTunes, Google Play, Amazon credits, every game session.
So every four hour on watsongo.co game page is published an object of real world, for example a bicycle, a book, a pot, a plant, a headphones, etc. This phase represents a challenge, a game session is composed by multiple challenges. For example, here is a challenge with #bicycle hashtag
you should find a bicycle closest match to the target object, where the target object is visual recognition class based on a set of images used to training the visual recognition machine learning. See my previous article here for details about how the Watson visual recognition works.
During every challenge you should found the object in a particular location, which can be initially a big city but in the next future we could use some points of interest (as Pokemon in Central Park 🙂 )
In this first stage as mobile app we are going to use Twitter on your favorite smartphone, here is a simple tweet example for the #macbook challenge:
Watson GO algorithm analyzes its own twitter timeline to assign a score to each user who participates in the game. The score takes into account how the photo is similar to images used to create the Visual Recognition class (from 0 to 10) plus a bonus if the location is equal to the challenge location (from 0 to 5).
On Watson Go game page you can check your score in real time:
And you can check the leader-board with the players with the best score:
Under the hood:
The game (alpha version) runs on IBM Bluemix cloud using the following services:
Few game-play mechanics used are still in alpha version, for example we are using the node-red-node-twitter and it can happen that the application exceed the api twitter limits, these errors will clear when the current 15 minute window passes. We are sorry if sometimes you do not see your score between the last tweets.
As mention we have used Bluemix services, in details the application is hosted on a node.js cloud foundry application. The Watson GO application consist of a:
- web-UI stateless based on html5 with bootstrap and angularjs
- node-red flows to manage the business logic and to provide the rest-APIs for web-UI
Join Alpha Testing!
we’re looking for alpha testers to help us to test Watson GO. If you meet a few prerequisites below and get excited about getting glimpses at per-released software, we want your help!
- a mobile device with a Twitter account 🙂
Thanks a lot and stay tuned!