They did this using a strategy developed with the help of a program researcher designed to probe the weaknesses of systems like KataGo. This victory seems of AlphaGO It’s just a small part of a broader Go renaissance that has seen human players get more creative since its milestone victory in 2016.
In a recent study published in the journal PNAS, researchers from the City University of Hong Kong and Yale discovered that human Go players have become less predictable in recent years. of New Scientist As he explains, the researchers came to this conclusion in the form of a test made during professional gaming between 1950 and 2021. 5.8 million more go They arrived at it by analyzing a dataset containing its movement. who can play the game program “superhuman” With the help of a Go AI and grading the quality of any move, they created a statistic called the “decision quality index” or DQI for short.
After the team assigned a DQI score to each action in their dataset, they saw professional game quality improve relatively little year-over-year before 2016. The team has a positive median annual maximum of 0.2. DQI detected the change.
Humans can be smarter than artificial intelligence!
In some years, the overall quality of the game has even dropped. With this, in 2018 Since the rise of superhuman AIs, median DQI values have changed by over 0.7. During the same period, professional players began to use newer strategies. In 2015, the rate of players forming an unprecedented game combination was 63%, while in 2018 this increased to 88%.
Team, “Our findings, superhuman AI “It shows that the development of our programs may have encouraged human players to move away from traditional strategies and explore new moves, which in turn may have improved their decision-making processes.”
This is an interesting change, but not exactly unintuitive if you think about it. Berkeley, California As University professor Stuart Russell told New Scientist, “it’s not surprising that players who train against machines tend to make more moves that the machines approve.”