This includes establishing likely tasks and needs associated with that future event to determine if the user has completed all of the things that need to be done before committing to that future event. This can include things like stopping at the gas station to refuel the night before a flight when the flight is due to take off at 6:30am. It should be noted that in other sections not included here, these machine learning and AI systems are constantly training to understand global patterns, as well as patterns specific to individuals. Chapter 72 [P]
the rediction module 264 may enter the current context, the previously determined future context, the current action and the expected or future action into the action rule data memory 270C and receive an indication of a degree of probability as to whether the user jewelry retouching service should be able to give up dry cleaning, they will have to attend the event the next day and still be able to attend the event the next day. …Action Rules Data Store 270C may generate a low degree of probability (e.g., less
than 50% probability) since the dry cleaner is closed the next day and therefore the user will not be able to recover their service dry cleaning for the event. Here's my favorite part and where it all comes together for Google. In this section, we see a user giving up dry cleaning, and the system predicted that it's for an event the user has the next day. Knowing that the dry cleaner is closed the next day, Google uses the corrective action by suggesting a nearby dry cleaner that will be open the next day.