Match-SRNN: Modeling the Recursive Matching Structure with Spatial RNN

By Shengxian Wan, Yanyan Lan, Jun Xu, Jiafeng Guo, Liang Pang, Xueqi Cheng. 2016

176 citations recorded by the source corpus.

Semantic matching, which aims to determine the matching degree between two texts, is a fundamental problem for many NLP applications. Recently, deep learning approach has been applied to this problem and significant improvements have been achieved. In this paper, we propose to view the generation of the global interaction between two texts as a recursive process: i.e. the interaction of two texts at each position is a composition of the interactions between their prefixes as well as the word level interaction at the current position. Based on this idea, we propose a novel deep architecture, namely Match-SRNN, to model the recursive matching structure. Firstly, a tensor is constructed to capture the word level interactions. Then a spatial RNN is applied to integrate the local interactions recursively, with importance determined by four types of gates. Finally, the matching score is calculated based on the global interaction. We show that, after degenerated to the exact matching scenario, Match-SRNN can approximate the dynamic programming process of longest common subsequence. Thus, there exists a clear interpretation for Match-SRNN. Our experiments on two semantic matching tasks showed the effectiveness of Match-SRNN, and its ability of visualizing the learned matching structure.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Natural Language Processing

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. Yixing Fan
5. Xueqi Cheng
6. Qingyao Ai
7. Andrew Yates
8. Jamie Callan

In Computer Vision and Pattern Recognition

1. Jiafeng Guo
2. Yixing Fan
3. W. Bruce Croft
4. Xueqi Cheng
5. Qingyao Ai
6. Wen Wang
7. Qian Chen
8. Liu Yang

In Natural Language Processing

1. Jiafeng Guo
2. W. Bruce Croft
3. Hamed Zamani
4. Yixing Fan
5. Xueqi Cheng
6. Wen Wang
7. Qingyao Ai
8. Liu Yang

Citations in the corpus

• A Deep Relevance Matching Model for Ad-hoc Retrieval (2016, background): “In [27], Wan et al. introduced Match-SRNN to model the recursive matching structure in the local interactions so that long-distance dependency between the interactions can be captured.”
• A Deep Look into Neural Ranking Models for Information Retrieval (2019, methodology): “A variety of neural ranking models [68, 18, 69, 70, 25] have been tested on these”
• An Introduction to Neural Information Retrieval (2018, unknown): “RNNs are popularly applied to sequences, but can also be useful for two (and higher) dimensional inputs [Wan et al., 2016].”
• Modeling Multi-turn Conversation with Deep Utterance Aggregation (2018, methodology): “…et al., 2015; Lowe et al., 2015), including TF-IDF, CNN, RNN, LSTM and biLSTM ; We also explore other advanced single-turn matching models, MV-LSTM (Wan et al., 2016), Match-LSTM (Wang and Jiang, 2015), Attentive-LSTM (Tan et al., 2015), Multi-Channels (Wu et al., 2017); These models concatenate…”
• Sequential Attention-based Network for Noetic End-to-End Response Selection (2019, methodology): “The second group of models consists of sequence-based matching models, which usually use the attention mechanism, including MV-LSTM (Wan et al. 2016), Matching-LSTM (Wang and Jiang 2016), Attentive-LSTM (Tan, Xi-ang, and Zhou 2015), and Multi-Channels (Wu et al. 2017).”
• Enhance word representation for out-of-vocabulary on Ubuntu dialogue corpus (2018, unknown): “…0.479 0.514 Multi-View (Zhou et al., 2016; Wu et al., 2017) 0.342 0.505 0.543 DL2R (Yan et al., 2016a; Wu et al., 2017) 0.330 0.488 0.527 MV-LSTM (Wan et al., 2016; Wu et al., 2017) 0.348 0.498 0.538 Match-LSTM (Wang & Jiang, 2015; Wu et al., 2017) 0.345 0.500 0.537 Attentive-LSTM (Tan et al.,…”
• BERT-INT: A BERT-based Interaction Model For Knowledge Graph Alignment (2020, methodology): “CNN [Hu et al. , 2014] and RNN [Wan et al. , 2016] are usually used as the aggregation function to extract the matching patterns from the similarity matrix between two sentences.”
• Modeling Diverse Relevance Patterns in Ad-hoc Retrieval (2018, methodology): “Specifically, we employ a spatial GRU model [34] for the relevance matching between a query and a passage, which can well capture semantic relations, proximities, and term importance.”