WOT2018 Xian Yunsen: Algorithms are everywhere in O2O search

[51CTO.com original article] Seven years of hard work, a gorgeous transformation. Since 2012, 15 summits have been held for 6 consecutive years, gathering a large number of technical experts, learning from a wide range of knowledge and making a breakthrough. The 2018 WOT Global Software and Operation Technology Summit will set sail on May 18, focusing on 12 core hotspots, bringing together 60 first-line experts from home and abroad to create a high-end technology feast!

On the second day, May 19, at the "Artificial Intelligence Technology Exploration" session at Venue A, Xian Yunsen, head of Meituan's search algorithm, will give a keynote speech on "Intelligent O2O Service Search." He told reporters that he will give an overall introduction to Meituan's search algorithm work, including pre-search recommendations, guidance during search, in-depth understanding of user intent, intelligent sorting, and joint optimization of multiple search scenarios. "I will also focus on two or three of our thoughtful and interesting work on algorithms to discuss with you."

Algorithms are everywhere

Xian Yunsen admitted that although Meituan's solution may not be the best, it has indeed done a lot of practical work in the field of intelligent search. "Through this conference, I will discuss with you the problems encountered in life services and share how Meituan analyzes and solves problems. I hope it will serve as a starting point and inspire everyone to actively think and explore technical solutions in complex business scenarios." Xian Yunsen said.

The reporter learned that service search has three characteristics, namely, scenario-based, localized (LBS), and focusing on transactions rather than just optimizing clicks (CTR). Xian Yunsen told reporters that algorithms are everywhere in O2O search. He described in detail the many application scenarios of the algorithm - starting from opening the APP, the search box will recommend the default search term, and after clicking the search box, there will be hot word recommendations, as well as subsequent suggestions, related searches and other guiding products; after submitting the search, user intent understanding and intelligent sorting will be deeply applied to the relevant algorithms of NLP and machine learning.

Exploring the algorithm behind food and marriage

According to Xian Yunsen, the large number of businesses and complex scenarios are a big challenge for O2O. Meituan has dozens of business lines, and the differences between different businesses are very large, which also puts high demands on algorithms. He took "food" and "marriage" as examples. When choosing "food", most people may pay more attention to distance and discounts, and the decision is light. Even if the choice fails, the worst thing is to choose a better restaurant next time; but "marriage" is different. It only happens once in a lifetime, the decision is heavy, and price and distance become less important.

Another challenge is that life services are highly scenario-based, requiring algorithms to provide more "intelligent" results. He also gave an example. For example, if a white-collar worker opens the Meituan app on Friday afternoon to look for hot springs, he will most likely go out to play and stay overnight; but if he does the same thing on Sunday night, he will most likely relax nearby. "Meituan Search has made many algorithm attempts for business and scenarios. We will build independent models for different businesses and optimize them in depth. The model will model scenarios and users, and understand user intentions in a more real-time manner."

Technology first is not desirable, business understanding is more important

Xian Yunsen also shared with reporters three lessons he learned from his practice, hoping to provide a reference for his peers.

First, he emphasized that as a technical manager, you should not neglect technology. You can focus on management work or develop personal management skills at certain stages, but it is dangerous to be out of touch with technology for a long time. Technology is developing very fast now, and you will fall behind if you don’t touch technology for 1-2 years. If a manager himself does not understand it, then he will definitely not do a good job of management.

Secondly, don't be deceived by your own experience. Especially for workers who have worked for many years and have rich experience, the more successful they have been in the past, the higher the risk of failure in the future. You must persist in learning, think hard, dare to break your own successful experience, and fully understand the business characteristics of the new job. Experience can be used as a reference, but it cannot be copied.

***He pointed out that business understanding is far more important than technology, and this sentence can be said a hundred times. Keeping up with the pace of technological development does not mean that technology can solve all problems. Analyzing problems and finding appropriate solutions is far more important than mastering advanced technology.

At the end of the interview, Xian Yunsen told reporters that the overall online rate of life services is not high at present. Taking food as an example, the percentage is still in the single digit, so life services will definitely continue to grow in the future. "Search is not a new technology, but there is still a lot of work to be done in the life service industry, and intelligent search is a strong demand."

On May 18-19, the world's most noteworthy IT technology feast will be held at Beijing JW Marriott Hotel. The 2018 WOT Global Software and Operation Technology Summit is an important platform for you to discover new ideas, explore new ideas, and expand your network.

At present, all our tickets are on sale. Please note that the earlier you buy your tickets, the bigger the discount! Communicate with KOLs up close and present a different "hero feast"!

Click on the official website for details: wot..com

20% off pre-sale, grab your tickets as soon as possible.

[51CTO original article, please indicate the original author and source as 51CTO.com when reprinting on partner sites]