学术讲座

报告标题：DynamicRetriever: A Pre-training Model-based IR System with Neither Sparse nor Dense Index

报告人简介：姚菁，硕士三年级

导师：文继荣，窦志成

研究方向：信息检索，个性化搜索，推荐

Abstract：Sparse retrieval and dense retrieval have become two mainstream methods for information retrieval. All these existing search methods follow a common paradigm, i.e. index-retrieve-rerank, a framework that has been around for decades. In this talk, I will introduce a new paradigm of search with neither spare nor dense index, which is a pre-training model-based IR system called DynamicRetriever. It hopes to embed the semantic information of documents on the Web into the model with the help of pre-training technology, so as to improve the effect of various downstream tasks of information retrieval. Then I will talk about its differences and advantages compared with existing methods.

报告标题：A Category-aware Multi-interest Model for Personalized Product Search

报告人简介：刘炯楠，博士一年级

导师：窦志成

研究方向：信息检索，个性化搜索

Abstract：Product search has been an important way for people to find products on online shopping platforms. Existing approaches in personalized product search mainly embed user preferences into one single vector. However, this simple strategy easily results in sub-optimal representations, failing to model and disentangle user’s multiple preferences. To overcome this problem, we proposed a category-aware multi-interest model to encode users as multiple preference embeddings to represent user-specific interests. Specifically, we also capture the category indications for each preference to indicate the distribution of categories it focuses on, which is derived from rich relations between users, products, and attributes. Based on these category indications, we develop a category attention mechanism to aggregate these various preference embeddings considering current queries and items as the user’s comprehensive representation. By this means, we can use this representation to calculate matching scores of retrieved items to determine whether they meet the user’s search intent. Besides, we introduce a homogenization regularization term to avoid the redundancy between user interests. Experimental results show that the proposed method significantly outperforms existing approaches.