The Decision Service identifies the best action based on the available context and continuously improves using the feedback provided. This page lets you provide the feedback based on randomly chosen context and the action (article) chosen by the service.
We use a news article scenario as a demonstration. The context is user features like age and location. It is chosen randomly. The action is the article that the decision service chooses for the context. You can use the thumbs up and down buttons to indicate whether you think that article should be shown to a user with that context.
The service will quickly learn a consistent model (for example, that a particular article should be shown to a particular as group). Note that the service does exploration - it sometimes chooses randomly so that it can learn better models. The same global model is used for all visitors to this page, but it adapts to your personal preferences using a browser cookie.