AI Can't Tell You Why It Did Something

The problem comes when the database and the engine go from )ach to oracle. It happens quite often that I will ask one of the students about a move from one of their games, and why he made it. If the move comes early on, the answer is almost always, "Because that's the nain line." That is, that's the theoretical move in the database, likely 5layed by many Grandmasters before. Sometimes the move isn't thery, but the student prepared it with the help of an engine, so the anwer is similar: "It's the best move." Maybe, yes, but, I always ask, why s it the best move? Why did all those Grandmasters play it? Why does 3 computer recommend it?

Then we often have a problem. Why? Because it is good. Why is it i? Answering that can take a lot of understanding and a lot of rerch. The openings have developed empirically over decades, sometimes over a century. If the bishop going to a certain square on move twelve in a specific variation is considered best, there's a whole story leading up to that moment, dozens or hundreds of games of trial and irror that went into establishing why that move precisely now.


Folksonomies: artificial intelligence

/art and entertainment/shows and events (0.674191)
/family and parenting/children (0.647601)
/society/dating (0.629280)

best move (0.677712 (:0.000000)), theoretical move (0.670636 (:0.000000)), specific variation (0.647070 (:0.000000)), whole story (0.641747 (:0.000000)), games of trial (0.576164 (:0.000000)), help of an engine (0.572209 (:0.000000)), move (0.557089 (:0.000000)), bishop (0.542959 (:0.000000)), students (0.537170 (:0.000000)), answer (0.533866 (:0.000000)), problem (0.533861 (:0.000000)), engine (0.533704 (:0.000000)), database (0.532140 (:0.000000)), lot of understanding (0.529961 (:0.000000)), games (0.527217 (:0.000000)), moment (0.526841 (:0.000000)), dozens (0.525867 (:0.000000)), openings (0.525390 (:0.000000)), student (0.518611 (:0.000000)), computer (0.513892 (:0.000000)), Grandmasters (0.513892 (:0.000000)), decades (0.509638 (:0.000000)), ach (0.509106 (:0.000000)), oracle (0.507511 (:0.000000)), century (0.507511 (:0.000000)), nain line (0.489812 (:0.000000)), lot of rerch (0.486087 (:0.000000)), irror (0.452228 (:0.000000)), anwer (0.451474 (:0.000000))

Grandmasters:Organization (0.849137 (:0.000000)), nain line:Facility (0.570386 (:0.000000)), AI:Organization (0.489014 (:0.000000))

Scientific method (0.903365): dbpedia_resource
Student (0.813047): dbpedia_resource
The Help (0.799208): dbpedia_resource
Theory (0.794733): dbpedia_resource
Answer (0.793471): dbpedia_resource
Ernest Hemingway (0.782000): dbpedia_resource

 Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins
Books, Brochures, and Chapters>Book:  Kasparov, Garry (201752), Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins, Retrieved on 2019-03-10
Folksonomies: artificial intelligence automation ai