Multi-Level Matching and Aggregation Network for Few-Shot Relation Classification

By Zhiquan Ye, Zhenhua Ling. 2019

148 citations recorded by the source corpus.

This paper presents a multi-level matching and aggregation network (MLMAN) for few-shot relation classification. Previous studies on this topic adopt prototypical networks, which calculate the embedding vector of a query instance and the prototype vector of the support set for each relation candidate independently. On the contrary, our proposed MLMAN model encodes the query instance and each support set in an interactive way by considering their matching information at both local and instance levels. The final class prototype for each support set is obtained by attentive aggregation over the representations of support instances, where the weights are calculated using the query instance. Experimental results demonstrate the effectiveness of our proposed methods, which achieve a new state-of-the-art performance on the FewRel dataset.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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. Philip S. Yu
2. Wanxiang Che
3. Ting-Wei Liu
4. Congying Xia
5. Haitao Zheng
6. Ning Ding
7. Hao Peng
8. Fei Huang

In Computer Vision and Pattern Recognition

1. Philip S. Yu
2. Wanxiang Che
3. Fei Huang
4. Hoang Nguyen
5. Haitao Zheng
6. Xindong Wu
7. Hao Peng
8. Lingfeng Zhong

Citations in the corpus

• A Two-phase Prototypical Network Model for Incremental Few-shot Relation Classification (2020, neutral): “They are listed as follows: Siamese (Koch et al., 2015), Proto (Han et al., 2018), LM-ProtoNet (Fan et al., 2019), HATT Proto (Gao et al., 2019a), MLMAN (Ye and Ling, 2019b), Proto(BERT), BERT-PAIR (Gao et al., 2019b).”
• Few-shot Slot Tagging with Collapsed Dependency Transfer and Label-enhanced Task-adaptive Projection Network (2020, background): “…including text classification (Sun et al., 2019; Geng et al., 2019; Yan et al., 2018; Yu et al., 2018), entity relation classification (Lv et al., 2019; Gao et al., 2019; Ye and Ling, 2019), and dialog act prediction (Vlasov et al., 2018 (Fader et al., 2013; Zhang et al., 2015; Liu et al., 2017).”
• Prototypical Representation Learning for Relation Extraction (2021, background): “…text in various specific scenarios, such as zero-few-shot scenarios (Levy et al., 2017; Han et al., 2018b; Gao et al., 2019; Soares et al., 2019; Ye & Ling, 2019) and open-domain scenarios (Banko et al., 2007; Fader et al., 2011; Mausam et al., 2012; Del Corro & Gemulla, 2013; Angeli et al.,…”
• Semantic Matching and Aggregation Network for Few-shot Intent Detection (2020, methodology): “The only exception is HAPN and MLMAN as they require local matching (i.e. word matching) modules.”
• A Comprehensive Survey on Automatic Knowledge Graph Construction (2023, background): “Another approach, called multi-Level matching and aggregation network (MLMAN) [207], aggregates the local and instance features by aggregating the support vectors with the query vector to match the correct long-tail class label for the query sample.”