Page - 82 - in The Future of Software Quality Assurance

82 R. Marselis faculties reliesupontheothers. Intelligence isacombinationofcognitiveskills and knowledgemadeevidentbybehaviors that are adaptive. Ability toLearn The ability to learn is the ability to comprehend, to understand, and to profit from experience.Howdoesan intelligentmachine learn? We see three levels of learning. The first level is rule-based learning. If a user uses certain options in a menu most frequently, the intelligent machine can order theoptionssuch that themostusedoptionsareon top.Thesecond level isbasedon gathering and interpreting data, and based on that learning about an environment. The third level is learning by observing behavior of others and imitating that behavior. Examplesof the levelsof learning: • At thefirst levelof learning, thinkofasatellite navigationsysteminacar, if you always turn off the automatic voice right after starting the system, the machine learns to startupwithout thevoiceactivated. • At the second level think of a roboticvacuum cleaner,by recording information about the layout it learns about the rooms that it cleans and therefore becomes better at avoidingobstaclesand reachingdifficult spots. • At the third level it’s about mimicking behavior, for example, a robot watches a YouTubevideoofbakingpancakesand thencopies thebehavior.Afterwatching severalvideos, the robotknowsall the tricksof the trade. Of course, the levelsof learningcanbecombinedby intelligentmachines. Improvisation Does it adapt to new situations? Improvisation is the power of the intelligent system to make right decisions in new situations. Such situations it has never experienced before require quick interpretation of new information and adjusting already existing behavior. Social robots must especially be able to adapt their behavioraccording to the informationcoming in, since social behaviordependson culture inspecificsmallgroups.Applyinglong-termchangeswill alsobe important fora robot to remain interestingor relevant for its environment. TransparencyofChoices Can a human involved understand how a machine comes to its decisions? An artificial intelligencesystemworks24/7and it takesa lotofdecisions.Therefore, it has tobe transparenthowanAIsystemtakesdecisions,basedonwhichdata inputs. And which data points are relevant and how are they weighted? In several use- cases, the decision-making is crucial. For example, when an Artificial Intelligent system calculates the insurance premium, it is important to investigate how this premium is calculated. Transparency also means predictability. It is important that robots respond as expected by the people who work with the robot. How well can the people involved foresee what (kind of) action the intelligent machine will take ina givensituation?This is thebasis forpropercollaboration(seenextparagraph).