Few-shot Relation Extraction via Bayesian Meta-learning on Relation Graphs

Meng Qu, Tianyu Gao, Louis-Pascal Xhonneux, Jian Tang
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:7867-7876, 2020.

Abstract

This paper studies few-shot relation extraction, which aims at predicting the relation for a pair of entities in a sentence by training with a few labeled examples in each relation. To more effectively generalize to new relations, in this paper we study the relationships between different relations and propose to leverage a global relation graph. We propose a novel Bayesian meta-learning approach to effectively learn the posterior distribution of the prototype vectors of relations, where the initial prior of the prototype vectors is parameterized with a graph neural network on the global relation graph. Moreover, to effectively optimize the posterior distribution of the prototype vectors, we propose to use the stochastic gradient Langevin dynamics, which is related to the MAML algorithm but is able to handle the uncertainty of the prototype vectors. The whole framework can be effectively and efficiently optimized in an end-to-end fashion. Experiments on two benchmark datasets prove the effectiveness of our proposed approach against competitive baselines in both the few-shot and zero-shot settings.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-qu20a, title = {Few-shot Relation Extraction via {B}ayesian Meta-learning on Relation Graphs}, author = {Qu, Meng and Gao, Tianyu and Xhonneux, Louis-Pascal and Tang, Jian}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {7867--7876}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/qu20a/qu20a.pdf}, url = {https://proceedings.mlr.press/v119/qu20a.html}, abstract = {This paper studies few-shot relation extraction, which aims at predicting the relation for a pair of entities in a sentence by training with a few labeled examples in each relation. To more effectively generalize to new relations, in this paper we study the relationships between different relations and propose to leverage a global relation graph. We propose a novel Bayesian meta-learning approach to effectively learn the posterior distribution of the prototype vectors of relations, where the initial prior of the prototype vectors is parameterized with a graph neural network on the global relation graph. Moreover, to effectively optimize the posterior distribution of the prototype vectors, we propose to use the stochastic gradient Langevin dynamics, which is related to the MAML algorithm but is able to handle the uncertainty of the prototype vectors. The whole framework can be effectively and efficiently optimized in an end-to-end fashion. Experiments on two benchmark datasets prove the effectiveness of our proposed approach against competitive baselines in both the few-shot and zero-shot settings.} }
Endnote
%0 Conference Paper %T Few-shot Relation Extraction via Bayesian Meta-learning on Relation Graphs %A Meng Qu %A Tianyu Gao %A Louis-Pascal Xhonneux %A Jian Tang %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-qu20a %I PMLR %P 7867--7876 %U https://proceedings.mlr.press/v119/qu20a.html %V 119 %X This paper studies few-shot relation extraction, which aims at predicting the relation for a pair of entities in a sentence by training with a few labeled examples in each relation. To more effectively generalize to new relations, in this paper we study the relationships between different relations and propose to leverage a global relation graph. We propose a novel Bayesian meta-learning approach to effectively learn the posterior distribution of the prototype vectors of relations, where the initial prior of the prototype vectors is parameterized with a graph neural network on the global relation graph. Moreover, to effectively optimize the posterior distribution of the prototype vectors, we propose to use the stochastic gradient Langevin dynamics, which is related to the MAML algorithm but is able to handle the uncertainty of the prototype vectors. The whole framework can be effectively and efficiently optimized in an end-to-end fashion. Experiments on two benchmark datasets prove the effectiveness of our proposed approach against competitive baselines in both the few-shot and zero-shot settings.
APA
Qu, M., Gao, T., Xhonneux, L. & Tang, J.. (2020). Few-shot Relation Extraction via Bayesian Meta-learning on Relation Graphs. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:7867-7876 Available from https://proceedings.mlr.press/v119/qu20a.html.

Related Material