Recursively Binary Modification Model for Nested Named Entity Recognition
Authors: Bing Li, Shifeng Liu, Yifang Sun, Wei Wang, Xiang Zhao8164-8171
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on four benchmark datasets demonstrate that our model outperforms state-of-the-art models in nested NER tasks, and delivers competitive results with state-of-the-art models in flat NER task, without relying on any extra annotations or NLP tools. |
| Researcher Affiliation | Academia | 1School of Computer Science and Engineering, University of New South Wales, Australia 2Dongguan University of Technology, China 3National University of Defence Technology, China |
| Pseudocode | No | The paper describes the model mathematically and textually but does not include any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any explicit statements about making the source code available or include a link to a code repository for the described methodology. |
| Open Datasets | Yes | We used the following four datasets, their detailed statistics summarized in Table 1. GENIA3 contains annotated entity mentions... ACE20054 is a multilingual training corpus... JNLPBA5 contains 36 fine-grained entity types... Co NLL03 (Tjong Kim Sang and De Meulder 2003) consists of newswire text... |
| Dataset Splits | Yes | We follow the settings of Ju et al. (2018) by keeping files from bc, bn, cts, nw and wl and randomly split them into training, development and testing sets with the ratio 8:1:1. |
| Hardware Specification | Yes | The Titan V used for this research was donated by the NVIDIA Corporation. |
| Software Dependencies | No | The paper mentions optimizers like Adam and various neural network architectures, but does not provide specific software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow versions, or specific library versions). |
| Experiment Setup | Yes | The default hyper-parameter settings were: the dimension of word-level and character-level embedding was 200 and 25, respectively. The dimension of hidden state of Bi LSTM was 200, and the undersampling rate m was set to 25. For optimization, we used Adam (Kinga and Adam 2015) with initial learning rate 0.001, weight-decay (L2) 1e 5, and the gradient clipping to 5; all other hyper-parameters were their default values. |