AI in the City
The story starts six years ago when British company DeepMind Technologies LTD. was founded. Four years later, Google acquired them.
[DeepMind] has created a neural network that learns how to play video games in a fashion similar to that of humans. — Wikipedia
Last year, AlphaGo a DeepMind project made history by defeating Go master Lee Sedol. AlphaGo is a narrow AI in which it performs a specific task very well. Google trained AlphaGo through competition with Go player Fan Hui.
The team building AlphaGo used the Monte Carlo Tree Search framework for training its decision making process.
The focus of Monte Carlo tree search is on the analysis of the most promising moves, expanding the search tree based on random sampling of the search space. — Wikipedia
This is a very powerful algorithm. Imagine teaching an AI to see a city as a board game. Projects like Vital Streets Project which look at equity around how do we better maintain and expand green infrastructure, biking paths, and connection to existing paths in Grand Rapids, MI could really benefit from developing an AI for learning the decision making. Dive in further and imagine sourcing data from multiple cities.
An AI could learn a point system based off some factors such as connecting existing paths, feedback from citizens, and most commute-friendly roads to have construction.
Metrics like positive or negative feedback could be sourced from citizens into a platform similar to Happy Maps.
Happy Maps might well contribute in changing the way engineering products are designed: often they are designed with the concept of efficiency in mind. But, being more efficient does not necessarily make us happier. — Happy Maps
The future is in engagement in community, neighborhoods, and cities. The infrastructure of these spaces could largely be developed by decision making advised by Artificial Intelligence. Now data is not the only tool to help make these decisions, computing power is standing along side. It is a very exciting time for mayors and citizens.