Incremental Event Detection via Knowledge Consolidation Networks

By Pengfei Cao, Yubo Chen, Jun Zhao, Taifeng Wang. 2020

52 citations recorded by the source corpus.

Conventional approaches to event detection usually require a fixed set of pre-defined event types. Such a requirement is often challenged in real-world applications, as new events continually occur. Due to huge computation cost and storage budge, it is infeasible to store all previous data and re-train the model with all previous data and new data, every time new events arrive. We formulate such challenging scenarios as incremental event detection, which requires a model to learn new classes incrementally without performance degradation on previous classes. However, existing incremental learning methods cannot handle semantic ambiguity and training data imbalance problems between old and new classes in the task of incremental event detection. In this paper, we propose a Knowledge Consolidation Network (KCN) to address the above issues. Specifically, we devise two components, prototype enhanced retrospection and hierarchical distillation, to mitigate the adverse effects of semantic ambiguity and class imbalance, respectively. Experimental results demonstrate the effectiveness of the proposed method, outperforming the state-of-the-art model by 19% and 13.4% of whole F1 score on ACE benchmark and TAC KBP benchmark, respectively.

Subfields: Artificial Intelligence

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. Jun Zhao
2. Kang Liu
3. Heng Ji
4. Pengfei Yu
5. P. Natarajan
6. Shizhu He
7. Zhuang Li
8. Yuan-Fang Li

Citations in the corpus

• Lifelong Intent Detection via Multi-Strategy Rebalancing (2021, background): “Due to the simplicity and effectiveness of replay-based methods, they become an excellent solution for lifelong learning in natural language processing scenarios [2, 9].”
• Lifelong Event Detection with Knowledge Transfer (2021, background): “KCN* (Cao et al., 2020) studies a different formulation of lifelong event detection.”
• Pretrained Language Model in Continual Learning: A Comparative Study (2022, methodology): “Maven (Wang et al., 2020) is a long-tailed event detection dataset with 163 classes, each of which is labelled at least 15 instances, reaching a total of 47,921 instances.”