Alphago and Alphastar
Robert H. Chen, Chelsea Chen · 2022
AlphaGo&s;s artificial neural network emulates the human brain&s;s network of neurons that are activated by input stimuli to form ideas by synaptic network connections producing “thought” patterns. While the best players like Lee Sedol and Ke Jie are always aggressively looking for and exploiting local fights, AlphaGo was expected to play a cold, computerized Deep Blue type top-down style to provoke and then engage in those fights to gain territory, but reinforced learning taught AlphaGo also to be є-greedy, often foregoing potential gains in a local fight to adventurously explore new board positions. However, exploratory forays that produce no positive gains in territory or stones can result in unnecessary losses of territory and stones by foregoing immediate exploitive fights, where it was believed that AlphaGo would be more effective, but, a supposedly a more coldly logical AlphaGo computer showed that it could be adventurous as well as meticulous.