Microsoft Research Connections Blog
Next at Microsoft
Social Media Collective
Windows on Theory
Posted by George Thomas Jr.
A new online Chinese riddle game is steeped in more than just tradition. In fact, the machine learning and artificial intelligence that fuels it derives from years of research that also helps drive Bing Search, Bing Translator, the Translator App for Windows Phone, and other products.
Launched in celebration of the Chinese New Year, Microsoft Chinese Character Riddle is based on the two-player game unique to Chinese traditional culture and part of the Chinese Lantern Festival. Developed by the Natural Language Computing Group in the Microsoft Research's Beijing lab, the game not only quickly returns an answer to a user's riddle, but also works in reverse: when a user enters a single Chinese character as the intended answer, the system generates several riddles from which to choose.
"These innovations typically embody the strategic thought of Natural Language Processing 2.0, which is to collect big data on the Internet, to automatically build AI models using statistical machine learning methods, and to involve users in the innovation process by quickly getting their on-line feedback." Says Dr. Ming Zhou, Group Leader for Natural Language Computing Group and Principal Researcher at Microsoft Research Asia. "Thus the riddle system will continue to improve."
Riddle: 两人相逢在异乡 (two persons meet in a foreign land)
Answer: 纵 (vertical, formed by characters in red)
The riddle system is actually built upon the success of Microsoft Couplet, launched in 2004, which resulted from a challenge by then lab director Harry Shum, now Executive Vice President, Technology and Research, who asked Zhou to build a computer that could generate Chinese couplets.
Couplets are poetry unique to traditional Chinese culture and date back thousands of years. Comprised of two phrases or sentences, the challenge of the couplet is to answer the initial statement with one that matches its rhythm, length and context. For example, a response to "Horses gallop to embrace blessings" could be "Goats dance to announce spring."
Zhou, embraced the challenge and designed a novel statistical machine translation-based approach which, over time and with an ever-expanding research team, resulted in the Microsoft Couplets website. The site was especially popular during Chinese New Year, with over 35,000 couplets created by users every day.
While improving the cultural image of Bing in China, the technologies used in couplets generation have also been used to improve statistical machine translation. The machine translation approach to couplets was introduced in the paper, Generating Chinese Couplets using a Statistical MT Approach (290 KB .pdf).
See also: Microsoft Translator Team Blog
Based on the decade-long, rich experiences of developing Microsoft Couplets, Zhou's group spent 4 months on the new riddle system, which involved a large dose of data mining to optimize statistical machine learning methods capable of converting numerous natural language descriptions into specific riddles in Chinese character. The webpage not only displays the riddle answer but also the reasoning process behind it, which could help shape how users think about the riddles.
The team's efforts in machine translation not only generated some 30 patents, but also significant advances in natural language processing, machine learning, and artificial intelligence — mainstream impact in the form of Bing Translator (Chinese-English), Bing Search, Bing Dictionary, Bing Input Method Editor (IME), the Translator App for Windows Phone (Chinese-English), sign language translator, and, circling back to Chinese culture, wildly popular intuitive web-based Chinese couplets and riddle systems. Now, Windows Phone apps for Couplets and Riddle are both available as free downloads.
"I believe we are the first to solve Chinese character riddle through a computer," Says Zhou, "With Microsoft Couplets and Microsoft Chinese Character Riddle, we hope to help modern Chinese people better inherit and carry on this tradition and cultural legacy through our AI models. It also can be an interesting tool to learn Chinese language."