A Concept Knowledge-Driven Keywords Retrieval Framework for Sponsored Search
This work addresses a specific problem in sponsored search advertising for companies like Baidu, focusing on improving keyword retrieval accuracy and revenue, but it appears incremental as it builds on existing data-driven methods by incorporating knowledge bases.
The paper tackles the problem of poor generalization in data-driven deep learning methods for retrieving synonymous keywords in sponsored search, particularly for entity-level long-tail instances, by proposing a knowledge-driven conceptual retrieval framework that uses concept tagging from a knowledge base; it shows effectiveness in offline and online experiments and has been applied to Baidu's system, yielding significant revenue improvement.
In sponsored search, retrieving synonymous keywords for exact match type is important for accurately targeted advertising. Data-driven deep learning-based method has been proposed to tackle this problem. An apparent disadvantage of this method is its poor generalization performance on entity-level long-tail instances, even though they might share similar concept-level patterns with frequent instances. With the help of a large knowledge base, we find that most commercial synonymous query-keyword pairs can be abstracted into meaningful conceptual patterns through concept tagging. Based on this fact, we propose a novel knowledge-driven conceptual retrieval framework to mitigate this problem, which consists of three parts: data conceptualization, matching via conceptual patterns and concept-augmented discrimination. Both offline and online experiments show that our method is very effective. This framework has been successfully applied to Baidu's sponsored search system, which yields a significant improvement in revenue.