Event Factuality Identification via Generative Adversarial Networks with Auxiliary Classification

Authors: Zhong Qian, Peifeng Li, Yue Zhang, Guodong Zhou, Qiaoming Zhu

IJCAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on Fact Bank show that our method significantly outperforms several stateof-the-art baselines, particularly on events with embedded sources, speculative and negative factuality values.
Researcher Affiliation Academia Zhong Qian1, Peifeng Li1, Yue Zhang2, Guodong Zhou1, Qiaoming Zhu1 1School of Computer Science and Technology, Soochow University, Suzhou, China 2Singapore University of Technology and Design, Singapore qianzhongqz@163.com, pfli@suda.edu.cn, yue zhang@sutd.edu.sg, {gdzhou, qmzhu}@suda.edu.cn
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code Yes The code of this paper is released at https: //github.com/qz011/ef_ac_gan.
Open Datasets Yes We evaluate our models on Fact Bank [Saur ı and Pustejovsky, 2009], which contains 3864 sentences and 13506 event factuality values.
Dataset Splits Yes For fair comparison, we perform 10-fold cross-validation on Fact Bank.
Hardware Specification No The paper does not specify any hardware details (e.g., GPU/CPU models, memory) used for running experiments.
Software Dependencies No The paper mentions software like Word2Vec and SGD, but does not provide specific version numbers for any libraries, frameworks, or programming languages.
Experiment Setup Yes For SIP detection we set the dimensions of the POS and hypernym embeddings as 50 and λ = 10 4. For event factuality identification, we set nlstm = 50 and dp = dt = 10.