Learning to Rank for Search Advertising

Ranking plays a critical role in both Web search and sponsored search. In sponsored search, the problem is even more complex and challenging because we need to consider multiple factors during the ranking process.

We are developing advanced ranking algorithms that can optimize relevance of ads, revenue of search engines, and satisfaction of advertisers. In this way, we can ensure the long-term revenue of search engines, instead of pursuing short-term return. Specifically, we study the true reasons for a user to click an ad, the factors that influence the satisfaction of an advertiser, and how they collectively influence the future revenue of search engines by economic theories.