Challenge Jan-2019

Identify Unfriendly Robots                                                                                      Solutions

A cyber police received the following information about Unfriendly and Friendly Robots:

The police asked its analysts to specify rules to identify if a robot is friendly or unfriendly, The rules were expressed in terms of any features you can see in robots, such as the shape of the head, the color of the jacket, the height, the color of their antennas, what they are holding in their hands, whether they are smiling or not, etc. The analysts quickly identified simple rules that succeeded for the robots from the above lists. Can you guess which rules were used?

However, when the police received more information about robots

the manually determined rules failed miserably. Please help the police to determine rules that can identify if any robot (from these lists or a new one) is friendly or unfriendly. You may use any combination of machine learning and business rules tools to come up with reliable rules.  Send your solutions to DecisionManagementCommunity@gmail.com.

Solutions:

P.S. The problem and images were borrowed from this 2008 research paper.

P.S.S. A really good solution should be able to adjust (learn) the rules as new information becomes available.