Pseudo-Relevance Feedback Based on Matrix Factorization

By Hamed Zamani, J. Dadashkarimi, Azadeh Shakery, W. Bruce Croft. 2016

62 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. W. Bruce Croft
2. Hamed Zamani
3. Jiafeng Guo
4. Jun Xu
5. Xueqi Cheng
6. Qingyao Ai
7. Yixing Fan
8. Jamie Callan

In Information Systems

1. Jun Xu
2. Hang Li
3. Hamed Zamani
4. W. Bruce Croft
5. Jiafeng Guo
6. Xiangnan He
7. Qingyao Ai
8. Xueqi Cheng

Citations in the corpus

• Relevance-based Word Embedding (2017, methodology): “These distributions are estimated based on the top k retrieved documents (in our experiments, we set k to 10) for each query. distributions can be estimated us-ing any PRF model, such as those proposed in [20, 36, 42, 44].”
• A Deep Look into Neural Ranking Models for Information Retrieval (2019, background): “[153] proposed RFMF, the first pseudo-relevance feedback model that learns latent factors from the top retrieved document.”
• Deep Learning for Matching in Search and Recommendation (2018, methodology): “Some search problems can be solved by using recommendation techniques (Zamani et al., 2016), and vice versa (Costa and Roda, 2011), on the basis of matching.”
• From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing (2018, unknown): “Following previous work on PRF [24, 29, 54, 59], we only keep the top t terms with the highest values in the updated query vector ® q∗.”
• Response Ranking with Deep Matching Networks and External Knowledge in Information-seeking Conversation Systems (2018, unknown): “Pseudo-relevance feedback (PRF) has been proven effective in improving the performance of many retrieval models [3, 12, 17, 25, 46, 47].”
• From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing (2018, unknown): “Following previous work on PRF [24, 29, 54, 59], we only keep the top t terms with the highest values in the updated query vector ® q∗.”