Group based Personalized Search by Integrating Search Behaviour and Friend Network

Zhou, Yujia; Dou, Zhicheng; Wei, Bingzheng; Xie, Rong; Wen, Ji-Rong

doi:10.1145/3404835.3462918

Cited by 25 publications

(10 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In social scenarios, the friend circle is a common relationship structure, such as classmates and relatives. One of the characteristics of the friend circle is that there is often one person at the core who is responsible for connecting people in the entire circle [52]. This feature provides us with a way to extract friend circles.…”

Section: Graph Partitionmentioning

confidence: 99%

“…To facilitate calculation, we propose to segment the graph into multiple subgraphs, while retaining as much information as possible. In social networks, the relationships between people usually depend on friend circles [10,13], and there is often a central person in each circle [52]. Based on this observation, we propose a graph partition algorithm which focuses on finding central nodes in the graph.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Socialformer: Social Network Inspired Long Document Modeling for Document Ranking

Zhou¹,

Dou²,

Yuan³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Utilizing pre-trained language models has achieved great success for neural document ranking. Limited by the computational and memory requirements, long document modeling becomes a critical issue. Recent works propose to modify the full attention matrix in Transformer by designing sparse attention patterns. However, most of them only focus on local connections of terms within a fixed-size window. How to build suitable remote connections between terms to better model document representation remains underexplored. In this paper, we propose the model Socialformer, which introduces the characteristics of social networks into designing sparse attention patterns for long document modeling in document ranking. Specifically, we consider several attention patterns to construct a graph like social networks. Endowed with the characteristic of social networks, most pairs of nodes in such a graph can reach with a short path while ensuring the sparsity. To facilitate efficient calculation, we segment the graph into multiple subgraphs to simulate friend circles in social scenarios. Experimental results confirm the effectiveness of our model on long document modeling. CCS CONCEPTS• Information systems → Retrieval models and ranking.

show abstract

Section: Graph Partitionmentioning

confidence: 99%

mentioning

confidence: 99%

Socialformer: Social Network Inspired Long Document Modeling for Document Ranking

Zhou¹,

Dou²,

Yuan³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In recent years, deep learning has been applied to user modeling for personalized search, which is effective in exploring the potential interests of users [12,19,23,40,[43][44][45]. An adaptive ranking model was devised in [28] for building dynamic user profiles.…”

Section: Related Work 21 Personalized Searchmentioning

confidence: 99%

PSSL: Self-supervised Learning for Personalized Search with Contrastive Sampling

Zhou¹,

Dou²,

Zhu³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Personalized search plays a crucial role in improving user search experience owing to its ability to build user profiles based on historical behaviors. Previous studies have made great progress in extracting personal signals from the query log and learning user representations. However, neural personalized search is extremely dependent on sufficient data to train the user model. Data sparsity is an inevitable challenge for existing methods to learn high-quality user representations. Moreover, the overemphasis on final ranking quality leads to rough data representations and impairs the generalizability of the model. To tackle these issues, we propose a Personalized Search framework with Self-supervised Learning (PSSL) to enhance data representations. Specifically, we adopt a contrastive sampling method to extract paired self-supervised information from sequences of user behaviors in query logs. Four auxiliary tasks are designed to pre-train the sentence encoder and the sequence encoder used in the ranking model. They are optimized by contrastive loss which aims to close the distance between similar user sequences, queries, and documents. Experimental results on two datasets demonstrate that our proposed model PSSL achieves state-of-the-art performance compared with existing baselines.

show abstract

“…In real-world applications, the queries issued by users are often short and ambiguous [28,41,60,62], such as the query "Apple" (Apple fruit or Apple company?). Thus, building an accurate encoding of the input query is difficult, which further leads to the poor quality of these ambiguous queries.…”

Section: Query Disambiguation Modeling (Qdm)mentioning

confidence: 99%

Pre-training for Ad-hoc Retrieval: Hyperlink is Also You Need

Ma,

Dou,

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Designing pre-training objectives that more closely resemble the downstream tasks for pre-trained language models can lead to better performance at the fine-tuning stage, especially in the ad-hoc retrieval area. Existing pre-training approaches tailored for IR tried to incorporate weak supervised signals, such as query-likelihood based sampling, to construct pseudo query-document pairs from the raw textual corpus. However, these signals rely heavily on the sampling method. For example, the query likelihood model may lead to much noise in the constructed pre-training data. In this paper, we propose to leverage the large-scale hyperlinks and anchor texts to pre-train the language model for ad-hoc retrieval. Since the anchor texts are created by webmasters and can usually summarize the target document, it can help to build more accurate and reliable pre-training samples than a specific algorithm. Considering different views of the downstream ad-hoc retrieval, we devise four pre-training tasks based on the hyperlinks. We then pre-train the Transformer model to predict the pair-wise preference, jointly with the Masked Language Model objective. Experimental results on two large-scale ad-hoc retrieval datasets show the significant improvement of our model compared with the existing methods. CCS CONCEPTS• Information systems → Retrieval models and ranking.

show abstract

Group based Personalized Search by Integrating Search Behaviour and Friend Network

Cited by 25 publications

References 36 publications

Socialformer: Social Network Inspired Long Document Modeling for Document Ranking

Socialformer: Social Network Inspired Long Document Modeling for Document Ranking

PSSL: Self-supervised Learning for Personalized Search with Contrastive Sampling

Pre-training for Ad-hoc Retrieval: Hyperlink is Also You Need

Contact Info

Product

Resources

About