[Global Science and Technology Report] March 29 news, a stop in the final guard of the boss - Sogou Wangzai robot, in a row down Harvard female tyrants Fengfeng Ting, and Wang Sicong mutual Ms. Red Miss sister look, IQ166 Talented young Liu Xingtu and “Nanca University Grassâ€, the undefeated god of war, such as the six-Lian Zhuang unbeaten god of war, Pluto, and other talented players, but yesterday “caught the greatest crisis in historyâ€.
It is understood that the current challenge of Wang Tsai is a graduate of Tsinghua University. The man from Inner Mongolia came up against him with his overbearing arrogance. Although he did not agree with Ji Qingye to finish the program after the match, he went to “eat meat and drinkâ€. Wang Tsai was not soft in the game and the start was blazing. In the next two cities, Kyrgyzstan’s advantage in the 3:0 start was directly reduced to 1 point. Afterwards, the scores of both sides bite each other and staggered. As the answer to the question has become more intense, the ratio is not correct or not, but it is the prejudgment and answer to the question.
In the answer to the question, Ji Qingyi and Wang Zai both answered questions correctly, but in the end Ji Zhe Mo was successful with a slight speed and slightly ahead of Wang Tsai. He scored a score of 7:4 and became In the end, the first person on the stage defeated Wang Tsai.
Since the launch of "One Stop to the End", how to defeat Wang Tsai has always been a matter of greatest concern to the players. The circulated on the Internet that "humans should try to answer the questions when the host does not ask questions" has also become a point of many players against Wang Tsai. consensus. Because Wang Zai does not rely on memorizing the game stage on the question bank, but on the basis of Sogou’s knowledge of the question answering system, after understanding the user's problems or information needs, he uses massive network information and a large-scale knowledge base to directly give answers. Humans can reason about the direction of the problem based on fragmentation information. To achieve this, the machine needs to overcome the difficulty in artificial intelligence—semantic understanding technology, how to seize the machine for semantic understanding, and the subtle time difference in reasoning and judgment. The key to victory, and this time Wang Zai was slightly inferior in time.
Of course, Wang Tsai, who is backed by Sogou’s knowledge system, is still learning. Sogou CEO Wang Xiaochuan said, “In a few years, there will be 20% or more problems. Sogou Lizhi can directly give an answer. In the future, it will reach 70%.†As Sogou’s investment in the smart question and answer field increases, Wang The ability to automatically answer all kinds of natural language problems will also continue to increase and grow even more intelligent.