Improving Low Resource Named Entity Recognition using Cross-lingual Knowledge Transfer

Authors: Xiaocheng Feng, Xiachong Feng, Bing Qin, Zhangyin Feng, Ting Liu

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on two low resource languages (Dutch and Spanish) demonstrate the effectiveness of these additional semantic representations (average 4.8% improvement). Moreover, on Chinese Onto Notes 4.0 dataset, our approach achieves an F-score of 83.07% with 2.91% absolute gain compared to the state-of-the-art systems.
Researcher Affiliation Academia Xiaocheng Feng, Xiachong Feng, Bing Qin , Zhangyin Feng, Ting Liu Harbin Institute of Technology, China {xcfeng, xiachongfeng, bqin, zyfeng, tliu}@ir.hit.edu.cn
Pseudocode No The paper includes architectural diagrams (Figure 2) and mathematical equations, but it does not provide any structured pseudocode or algorithm blocks.
Open Source Code Yes Our code is available at: https://github.com/scir-code/lrner.
Open Datasets Yes Table 2: # of sentences. Languages Dataset Train Dev Test Dutch Co NLL-2002 15520 2822 5077 Spanish Co NLL-2002 8323 1914 1517 Chinese Ontonotes 4.0 22761 3903 2730. Also mentions 'Co NLL: https://github.com/synalp/NER/tree/master/corpus/' and 'Ontonotes: https://catalog.ldc.upenn.edu/ldc2011t03'.
Dataset Splits Yes Table 2: # of sentences. Languages Dataset Train Dev Test Dutch Co NLL-2002 15520 2822 5077 Spanish Co NLL-2002 8323 1914 1517 Chinese Ontonotes 4.0 22761 3903 2730
Hardware Specification No The paper does not provide any specific details about the hardware used for running the experiments (e.g., GPU models, CPU types, or memory specifications).
Software Dependencies No The paper mentions 'LSTM-CRF model is originally introduced by [Huang et al., 2015]' and refers to 'LSTM-CRF framework [Lample et al., 2016]' and states 'Our code is available at: https://github.com/scir-code/lrner.' but it does not list specific software dependencies with their version numbers (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes We randomize other parameters with uniform distribution U(0.01, 0.01), and set the learning rate as 0.01.