Knowledge Enhanced Event Causality Identification with Mention Masking Generalizations

Authors: Jian Liu, Yubo Chen, Jun Zhao

IJCAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In experiments, we evaluate our model on three benchmark datasets and show our model outperforms previous methods by a significant margin. Moreover, we perform 1) cross-topic adaptation, 2) exploiting unseen predicates, and 3) crosstask adaptation to evaluate the generalization ability of our model. Experimental results show that our model demonstrates a definite advantage over previous methods.
Researcher Affiliation Academia Jian Liu1,2 , Yubo Chen1,2 and Jun Zhao1,2 1 National Laboratory of Pattern Recognition, Institute of Automation Chinese Academy of Sciences, Beijing, 100190, China 2 School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, 100049, China
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes In our implementations1, both the knowledge aware reasoner and the mention masking reasoner are implemented as BERT-Large architecture, which has 24layer, 1024-hidden, and 16-heads. We use CONCEPTNET 5.0 as the KB. Regarding hyper-parameters, the batch size is set as 10, and the learning rate is initialized as 5 10 5 with a linear decay. We also adopt a negative sampling rate of 0.5 for training, owing to the sparseness of positive examples. 1https://github.com/jianliu-ml/Event Causality Identification
Open Datasets Yes Our experiments are conducted on three benchmark datasets, including: a) Event Strory Line [Caselli and Vossen, 2017], which contains 258 documents in 22 topics, 5,334 events in total, and 1,770 of 7,805 event pairs are causally related; b) Causal-Time Bank [Mirza et al., 2014], which contains 184 documents, 6,813 events, and 318 of 7,608 event pairs are causally related. c) Event Causality [Do et al., 2011; Ning et al., 2018], which contains 25 documents, 1,134 events, and 414 of 887 event pairs are causally related.
Dataset Splits Yes For Event Story Line, where we use the last two topics as development set, and conduct a 5-fold cross-validation on the rest 20 topics, as suggested by [Gao et al., 2019]. For Causal-Time Bank, where we adopt two settings: 1) 10-fold crossvalidation (CV) as in [Mirza, 2014a]...
Hardware Specification No The paper does not provide specific hardware details such as GPU or CPU models, processor types, or memory amounts used for running its experiments. It mentions 'BERT-Large architecture' but no underlying hardware.
Software Dependencies Yes In our implementations1, both the knowledge aware reasoner and the mention masking reasoner are implemented as BERT-Large architecture... We use CONCEPTNET 5.0 as the KB.
Experiment Setup Yes Regarding hyper-parameters, the batch size is set as 10, and the learning rate is initialized as 5 10 5 with a linear decay. We also adopt a negative sampling rate of 0.5 for training, owing to the sparseness of positive examples.