Checkmated through artificial intelligence

Chess playing has no immediate practical use. However it has been since the 19th century a high valuable experiment field for developing an “artificial intelligence”.

And since the development of the first chess playing computer by Alex Bernstein based on the  pioneering theoretical works from Alan Turing and later Claude Shannon  until the defeat of the world master Garry Kasparov against the chess engine Deep Blue in 1997, the progress have been immense.

These accomplishments have but nothing to do with an “artificial intelligence” but are singly the result of powerful hardware and better search algorithms. There are the fruits of the scientists ingenuity that have fed the program with ever more sophisticated human heuristic for position evaluation, inherited from hundreds of years of practice.

The new engine “alpha-zero” based on machine learning developed by the Deepmind company is incredibly surpassing the best deep Blue successors. Unlike the previous engines, AlphaZero isn’t applying pre-programmed human knowledge and calculating through billions of positions to generate moves – It is rather learning like us although in a different way : It finds the best way to play by training against itself though millions of games. This is a self-taught machine.

The following position against one the strongest “classical” chess engine "Stockfish" illustrates its “intuitive” play style: Alpha-Zero do have white and the black opponent though big material advantage is like entrapped in a cage.

The Alpha-Zero developers have used a subset of machine learning, deep learning to reach their milestones. Deep learning is one of the analytical technique that has reactivated the hopes but also fears on the AI technology in the recent years. Hidebound disciplines like education and medicine will also for sure be concerned by the improved results promised by AI analysis.

In our “Business and Science encounters” around the chess board, we want to discuss about the possibility but also the limitations of these new technology from the chess game into the daily life.

Look at the programm here

Organized together with: