| { |
| "paper_id": "U17-1013", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T03:11:22.000748Z" |
| }, |
| "title": "From Word Segmentation to POS Tagging for Vietnamese", |
| "authors": [ |
| { |
| "first": "Dat", |
| "middle": [ |
| "Quoc" |
| ], |
| "last": "Nguyen", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Macquarie University", |
| "location": { |
| "country": "Australia" |
| } |
| }, |
| "email": "dat.nguyen@mq.edu.au" |
| }, |
| { |
| "first": "Thanh", |
| "middle": [], |
| "last": "Vu", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Newcastle University", |
| "location": { |
| "country": "United Kingdom" |
| } |
| }, |
| "email": "thanh.vu@newcastle.ac.uk" |
| }, |
| { |
| "first": "Dai", |
| "middle": [ |
| "Quoc" |
| ], |
| "last": "Nguyen", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Deakin University", |
| "location": { |
| "country": "Australia" |
| } |
| }, |
| "email": "dai.nguyen@deakin.edu.au" |
| }, |
| { |
| "first": "Mark", |
| "middle": [], |
| "last": "Dras", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Macquarie University", |
| "location": { |
| "country": "Australia" |
| } |
| }, |
| "email": "mark.dras@mq.edu.au" |
| }, |
| { |
| "first": "Mark", |
| "middle": [], |
| "last": "Johnson", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Macquarie University", |
| "location": { |
| "country": "Australia" |
| } |
| }, |
| "email": "mark.johnson@mq.edu.au" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "This paper presents an empirical comparison of two strategies for Vietnamese Part-of-Speech (POS) tagging from unsegmented text: (i) a pipeline strategy where we consider the output of a word segmenter as the input of a POS tagger, and (ii) a joint strategy where we predict a combined segmentation and POS tag for each syllable. We also make a comparison between state-of-the-art (SOTA) featurebased and neural network-based models. On the benchmark Vietnamese treebank (Nguyen et al., 2009), experimental results show that the pipeline strategy produces better scores of POS tagging from unsegmented text than the joint strategy, and the highest accuracy is obtained by using a feature-based model.", |
| "pdf_parse": { |
| "paper_id": "U17-1013", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "This paper presents an empirical comparison of two strategies for Vietnamese Part-of-Speech (POS) tagging from unsegmented text: (i) a pipeline strategy where we consider the output of a word segmenter as the input of a POS tagger, and (ii) a joint strategy where we predict a combined segmentation and POS tag for each syllable. We also make a comparison between state-of-the-art (SOTA) featurebased and neural network-based models. On the benchmark Vietnamese treebank (Nguyen et al., 2009), experimental results show that the pipeline strategy produces better scores of POS tagging from unsegmented text than the joint strategy, and the highest accuracy is obtained by using a feature-based model.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "POS tagging is one of the most fundamental natural language processing (NLP) tasks. In English where white space is a strong indicator of word boundaries, POS tagging is an important first step towards many other NLP tasks. However, white space when written in Vietnamese is also used to separate syllables that constitute words. So for Vietnamese NLP, word segmentation is referred to as the key first step (Dien et al., 2001) .", |
| "cite_spans": [ |
| { |
| "start": 408, |
| "end": 427, |
| "text": "(Dien et al., 2001)", |
| "ref_id": "BIBREF3" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "When applying POS tagging to real-world Vietnamese text where gold word-segmentation is not available, the pipeline strategy is to first segment the text by using a word segmenter, and then feed the word-segmented text-which is the output of the word segmenter-as the input to a POS tagger. For example, given a written text \"thu\u220f thu nh\u2122p c\u00e1 nh\u00e2n\" (individual c\u00e1_nh\u00e2n income thu_nh\u2122p tax thu\u220f ) consisting of 5 syllables, the word seg-menter returns a two-word phrase \"thu\u220f_thu_nh\u2122p c\u00e1_nh\u00e2n.\" 1 Then given the input segmented text \"thu\u220f_thu_nh\u2122p c\u00e1_nh\u00e2n\", the POS tagger returns \"thu\u220f_thu_nh\u2122p/N c\u00e1_nh\u00e2n/N.\"", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "A class of approaches to POS tagging from unsegmented text that has been actively explored in other languages, such as in Chinese and Japanese, is joint word segmentation and POS tagging (Zhang and Clark, 2008) . A possible joint strategy is to assign a combined segmentation and POS tag to each syllable (Kruengkrai et al., 2009) . For example, given the input text \"thu\u220f thu nh\u2122p c\u00e1 nh\u00e2n\", the joint strategy would produce \"thu\u220f/B-N thu/I-N nh\u2122p/I-N c\u00e1/B-N nh\u00e2n/I-N\", where B refers to the beginning of a word and I refers to the inside of a word. Shao et al. (2017) showed that this joint strategy gives SOTA results for Chinese POS tagging by utilizing a BiLSTM-CNN-CRF model (Ma and Hovy, 2016) .", |
| "cite_spans": [ |
| { |
| "start": 187, |
| "end": 210, |
| "text": "(Zhang and Clark, 2008)", |
| "ref_id": "BIBREF39" |
| }, |
| { |
| "start": 305, |
| "end": 330, |
| "text": "(Kruengkrai et al., 2009)", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 550, |
| "end": 568, |
| "text": "Shao et al. (2017)", |
| "ref_id": "BIBREF32" |
| }, |
| { |
| "start": 680, |
| "end": 699, |
| "text": "(Ma and Hovy, 2016)", |
| "ref_id": "BIBREF12" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "In this paper, we present the first empirical study comparing the joint and pipeline strategies for Vietnamese POS tagging from unsegmented text. In addition, we make a comparison between SOTA feature-based and neural networkbased models, which, to the best of our knowledge, has not done in any prior work on Vietnamese. On the benchmark Vietnamese treebank (Nguyen et al., 2009) , we show that the pipeline strategy produces better scores than the joint strategy. We also show that the highest tagging accuracy is obtained by using a traditional feature-based model rather than neural network-based models. Nguyen et al. (2006) , Dinh and Vu (2006) and Tran et al. (2010) considered the Vietnamese word segmentation task as a sequence labeling task, using either a CRF, SVM or MaxEnt model to assign each syllable a segmentation tag such as B or I. In addition, Le et al. (2008) , Pham et al. (2009) and Tran et al. (2012) used the maximum matching method (NanYuan and YanBin, 1991) to generate all possible segmentations for each input sentence; then to select the best segmentation, Le et al. (2008) and Tran et al. (2012) applied ngram language model while Pham et al. (2009) employed POS information from an external POS tagger. Later, Liu and Lin (2014) and Nguyen and Le (2016) proposed approaches based on pointwise prediction, where a binary classifier is trained to identify whether or not there is a word boundary at each point between two syllables. Furthermore, Nguyen et al. (2017b) proposed a rule-based approach which gets the highest results to date in terms of both segmentation accuracy and speed.", |
| "cite_spans": [ |
| { |
| "start": 359, |
| "end": 380, |
| "text": "(Nguyen et al., 2009)", |
| "ref_id": "BIBREF26" |
| }, |
| { |
| "start": 609, |
| "end": 629, |
| "text": "Nguyen et al. (2006)", |
| "ref_id": "BIBREF17" |
| }, |
| { |
| "start": 632, |
| "end": 650, |
| "text": "Dinh and Vu (2006)", |
| "ref_id": "BIBREF4" |
| }, |
| { |
| "start": 655, |
| "end": 673, |
| "text": "Tran et al. (2010)", |
| "ref_id": "BIBREF37" |
| }, |
| { |
| "start": 864, |
| "end": 880, |
| "text": "Le et al. (2008)", |
| "ref_id": "BIBREF9" |
| }, |
| { |
| "start": 883, |
| "end": 901, |
| "text": "Pham et al. (2009)", |
| "ref_id": "BIBREF29" |
| }, |
| { |
| "start": 906, |
| "end": 924, |
| "text": "Tran et al. (2012)", |
| "ref_id": "BIBREF35" |
| }, |
| { |
| "start": 958, |
| "end": 984, |
| "text": "(NanYuan and YanBin, 1991)", |
| "ref_id": "BIBREF15" |
| }, |
| { |
| "start": 1087, |
| "end": 1103, |
| "text": "Le et al. (2008)", |
| "ref_id": "BIBREF9" |
| }, |
| { |
| "start": 1108, |
| "end": 1126, |
| "text": "Tran et al. (2012)", |
| "ref_id": "BIBREF35" |
| }, |
| { |
| "start": 1162, |
| "end": 1180, |
| "text": "Pham et al. (2009)", |
| "ref_id": "BIBREF29" |
| }, |
| { |
| "start": 1242, |
| "end": 1260, |
| "text": "Liu and Lin (2014)", |
| "ref_id": "BIBREF11" |
| }, |
| { |
| "start": 1265, |
| "end": 1285, |
| "text": "Nguyen and Le (2016)", |
| "ref_id": "BIBREF27" |
| }, |
| { |
| "start": 1476, |
| "end": 1497, |
| "text": "Nguyen et al. (2017b)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Regarding Vietnamese POS tagging, Dien and Kiem (2003) projected POS annotations from English to Vietnamese via a bilingual corpus of word alignments. As a standard sequence labeling task, previous research has applied the CRF, SVM or MaxEnt model to assign each word a POS tag (Nghiem et al., 2008; Le-Hong et al., 2010; Tran et al., 2010; Bach et al., 2013) . In addition, Nguyen et al. (2011) proposed a rule-based approach to automatically construct transformation rules for POS tagging in the form of a Ripple Down Rules tree (Compton and Jansen, 1990) , leading to a development of the RDRPOSTagger (Nguyen et al., 2014a) which was the best system for the POS tagging shared task at the 2013 Vietnamese Language and Speech Processing (VLSP) workshop. Nguyen et al. (2016a) and Nguyen et al. (2016b) later showed that SOTA accuracies at 94+% in the Vietnamese POS tagging task are obtained by simply retraining existing English POS taggers on Vietnamese data, showing that the MarMoT tagger (Mueller et al., 2013) and the Stanford POS tagger (Toutanova et al., 2003) obtain higher accuracies than RDRPOSTagger. Nguyen et al. (2016a) also showed that a simple lexicon-based approach assigning each word by its most probable POS tag gains a promising accuracy at 91%. Note that both Nguyen et al. (2016a) and Nguyen et al. (2016b) did not experiment with neural network models. Pham et al. (2017) recently applied the BiLSTM-CNN-CRF (Ma and Hovy, 2016) for Vietnamese POS tagging, however, they did not experiment with SOTA feature-based models.", |
| "cite_spans": [ |
| { |
| "start": 34, |
| "end": 54, |
| "text": "Dien and Kiem (2003)", |
| "ref_id": "BIBREF2" |
| }, |
| { |
| "start": 278, |
| "end": 299, |
| "text": "(Nghiem et al., 2008;", |
| "ref_id": "BIBREF16" |
| }, |
| { |
| "start": 300, |
| "end": 321, |
| "text": "Le-Hong et al., 2010;", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 322, |
| "end": 340, |
| "text": "Tran et al., 2010;", |
| "ref_id": "BIBREF37" |
| }, |
| { |
| "start": 341, |
| "end": 359, |
| "text": "Bach et al., 2013)", |
| "ref_id": "BIBREF0" |
| }, |
| { |
| "start": 531, |
| "end": 557, |
| "text": "(Compton and Jansen, 1990)", |
| "ref_id": "BIBREF1" |
| }, |
| { |
| "start": 605, |
| "end": 627, |
| "text": "(Nguyen et al., 2014a)", |
| "ref_id": null |
| }, |
| { |
| "start": 757, |
| "end": 778, |
| "text": "Nguyen et al. (2016a)", |
| "ref_id": null |
| }, |
| { |
| "start": 783, |
| "end": 804, |
| "text": "Nguyen et al. (2016b)", |
| "ref_id": "BIBREF28" |
| }, |
| { |
| "start": 996, |
| "end": 1018, |
| "text": "(Mueller et al., 2013)", |
| "ref_id": "BIBREF14" |
| }, |
| { |
| "start": 1047, |
| "end": 1071, |
| "text": "(Toutanova et al., 2003)", |
| "ref_id": "BIBREF34" |
| }, |
| { |
| "start": 1116, |
| "end": 1137, |
| "text": "Nguyen et al. (2016a)", |
| "ref_id": null |
| }, |
| { |
| "start": 1286, |
| "end": 1307, |
| "text": "Nguyen et al. (2016a)", |
| "ref_id": null |
| }, |
| { |
| "start": 1312, |
| "end": 1333, |
| "text": "Nguyen et al. (2016b)", |
| "ref_id": "BIBREF28" |
| }, |
| { |
| "start": 1381, |
| "end": 1399, |
| "text": "Pham et al. (2017)", |
| "ref_id": "BIBREF30" |
| }, |
| { |
| "start": 1436, |
| "end": 1455, |
| "text": "(Ma and Hovy, 2016)", |
| "ref_id": "BIBREF12" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "POS tagging", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "Previously, only Takahashi and Yamamoto (2016) carried out joint word segmentation and POS tagging for Vietnamese, to predicting a combined segmentation and POS tag to each syllable. In particular, Takahashi and Yamamoto (2016) experimented with traditional SVM-and CRFbased toolkits on a dataset of about 7k sentences and reported results of joint prediction only, i.e., they did not compare to the pipeline strategy. The CoNLL 2017 shared task on Universal Dependencies (UD) parsing from raw text (Zeman et al., 2017) provided some results to the pipeline strategy from word segmentation to POS tagging, however, the Vietnamese dataset in the UD project is very small, consisting of 1,400 training sentences. Furthermore, Nguyen et al. (2017a) provided a pre-trained jPTDP model for joint POS tagging and dependency parsing for Vietnamese, 2 which obtains a tagging accuracy at 93.0%, a UAS score at 77.7% and a LAS score at 69.5% when evaluated on the Vietnamese dependency treebank VnDT of 10k sentences (Nguyen et al., 2014b) .", |
| "cite_spans": [ |
| { |
| "start": 198, |
| "end": 227, |
| "text": "Takahashi and Yamamoto (2016)", |
| "ref_id": "BIBREF33" |
| }, |
| { |
| "start": 499, |
| "end": 519, |
| "text": "(Zeman et al., 2017)", |
| "ref_id": "BIBREF38" |
| }, |
| { |
| "start": 1008, |
| "end": 1030, |
| "text": "(Nguyen et al., 2014b)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "POS tagging", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "We compare the joint word segmentation and POS tagging strategy to the pipeline strategy on the benchmark Vietnamese treebank (Nguyen et al., 2009) using well-known POS tagging models.", |
| "cite_spans": [ |
| { |
| "start": 126, |
| "end": 147, |
| "text": "(Nguyen et al., 2009)", |
| "ref_id": "BIBREF26" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Experimental methodology", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Following Kruengkrai et al. (2009) , Takahashi and Yamamoto (2016) and Shao et al. (2017) , we formalize the joint word segmentation and POS tagging problem for Vietnamese as a sequence labeling task to assigning a combined segmentation and POS tag to each syllable. For example, given a manually POS-annotated training corpus \"Cu\u00cec/Nc i\u2211u_tra/V d\u02dc\u00ccng_nh\u02dc/X kh\u00f4ng/R ti\u220fn_tri\u222bn/V ./CH\" 'The investigation seems to be making no progress', we transform this corpus into a syllable-based representation as follows:", |
| "cite_spans": [ |
| { |
| "start": 10, |
| "end": 34, |
| "text": "Kruengkrai et al. (2009)", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 37, |
| "end": 66, |
| "text": "Takahashi and Yamamoto (2016)", |
| "ref_id": "BIBREF33" |
| }, |
| { |
| "start": 71, |
| "end": 89, |
| "text": "Shao et al. (2017)", |
| "ref_id": "BIBREF32" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Joint segmentation and POS tagging", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "\"Cu\u00cec/B-Nc i\u2211u/B-V tra/I-V d\u02dc\u00ccng/B-X nh\u02dc/I-X kh\u00f4ng/B-R ti\u220fn/B-V tri\u222bn/I-V ./B-CH\",", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Joint segmentation and POS tagging", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "where segmentation tags B and I denote beginning and in-2 https://drive.google.com/drive/ folders/0B5eBgc8jrKtpUmhhSmtFLWdrTzQ side of a word, respectively, while Nc, V, X, R and CH are POS tags. Then we train sequence labeling models on the syllable-based transformed corpus.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Joint segmentation and POS tagging", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "The Vietnamese treebank (Nguyen et al., 2009) is the largest annotated corpus for Vietnamese, providing a set of 27,870 manually POS-annotated sentences for training and development (about 23 words per sentence on average) and a test set of 2120 manually POS-annotated sentences (about 31 words per sentence). 3 From the set of 27,870 sentences, we use the first 27k sentences for training and the last 870 sentences for development.", |
| "cite_spans": [ |
| { |
| "start": 24, |
| "end": 45, |
| "text": "(Nguyen et al., 2009)", |
| "ref_id": "BIBREF26" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Dataset", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "For both joint and pipeline strategies, we use the following models:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Models", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "\u2022 RDRPOSTagger (Nguyen et al., 2014a ) is a transformation rule-based learning model which obtained the highest accuracy at the VLSP 2013 POS tagging shared task. 4", |
| "cite_spans": [ |
| { |
| "start": 15, |
| "end": 36, |
| "text": "(Nguyen et al., 2014a", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Models", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "\u2022 MarMoT (Mueller et al., 2013 ) is a generic CRF framework and a SOTA POS and morphological tagger. 5", |
| "cite_spans": [ |
| { |
| "start": 9, |
| "end": 30, |
| "text": "(Mueller et al., 2013", |
| "ref_id": "BIBREF14" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Models", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "\u2022 BiLSTM-CRF (Huang et al., 2015 ) is a sequence labeling model which extends the BiLSTM model with a CRF layer.", |
| "cite_spans": [ |
| { |
| "start": 13, |
| "end": 32, |
| "text": "(Huang et al., 2015", |
| "ref_id": "BIBREF6" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Models", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "\u2022 BiLSTM-CRF + CNN-char, i.e. BiLSTM-CNN-CRF, is an extension of the BiLSTM-CRF, using CNN to derive character-based representations (Ma and Hovy, 2016) .", |
| "cite_spans": [ |
| { |
| "start": 133, |
| "end": 152, |
| "text": "(Ma and Hovy, 2016)", |
| "ref_id": "BIBREF12" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Models", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "\u2022 BiLSTM-CRF + LSTM-char is another extension of the BiLSTM-CRF, using BiLSTM to derive the character-based representations (Lample et al., 2016) .", |
| "cite_spans": [ |
| { |
| "start": 124, |
| "end": 145, |
| "text": "(Lample et al., 2016)", |
| "ref_id": "BIBREF8" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Models", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "Here, for the pipeline strategy, we train these models to predict POS tags with respect to (w.r.t.) gold word segmentation. In addition, we also retrain the fast and accurate Vietnamese word segmenter RDRsegmenter (Nguyen et al., 2017b) Table 1 : Optimal number of LSTM units.", |
| "cite_spans": [ |
| { |
| "start": 214, |
| "end": 236, |
| "text": "(Nguyen et al., 2017b)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 237, |
| "end": 244, |
| "text": "Table 1", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Models", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "We use the original pure Java implementations of RDRPOSTagger and MarMoT with default hyperparameter settings in our experiments. Instead of using implementations independently provided by authors of BiLSTM-CRF, BiLSTM-CRF + CNNchar 7 and BiLSTM-CRF + LSTM-char, we use a reimplementation which is optimized for performance of all these models from Reimers and Gurevych (2017) . 8 For three BiLSTM-CRF-based models, we use default hyper-parameters provided by Reimers and Gurevych (2017) with the following exceptions: we use a dropout rate at 0.5 (Ma and Hovy, 2016) with the frequency threshold of 5 for unknown word and syllable types. We initialize word and syllable embeddings with 100-dimensional pretrained embeddings, 9 then learn them together with other model parameters during training by using Nadam (Dozat, 2016) . For training, we run for 100 epochs. We perform a grid search of hyperparameters to select the number of BiLSTM layers from {1, 2, 3} and the number of LSTM units in each layer from {50, 100, 150, 200, 250, 300}. Early stopping is applied when no performance improvement on the development set is obtained after 5 contiguous epochs. For both pipeline and joint strategies, we find the highest performance on the development set is when using two stacked BiLSTM layers. Table 1 presents the optimal number of LSTM units.", |
| "cite_spans": [ |
| { |
| "start": 349, |
| "end": 376, |
| "text": "Reimers and Gurevych (2017)", |
| "ref_id": "BIBREF31" |
| }, |
| { |
| "start": 379, |
| "end": 380, |
| "text": "8", |
| "ref_id": null |
| }, |
| { |
| "start": 460, |
| "end": 487, |
| "text": "Reimers and Gurevych (2017)", |
| "ref_id": "BIBREF31" |
| }, |
| { |
| "start": 548, |
| "end": 567, |
| "text": "(Ma and Hovy, 2016)", |
| "ref_id": "BIBREF12" |
| }, |
| { |
| "start": 812, |
| "end": 825, |
| "text": "(Dozat, 2016)", |
| "ref_id": "BIBREF5" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 1297, |
| "end": 1304, |
| "text": "Table 1", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Implementation details", |
| "sec_num": "3.4" |
| }, |
| { |
| "text": "Here the performance is evaluated by F1 score, based on the number of correctly segmented and tagged words (Zhang and Clark, 2008) . In the case of gold word segmentation, F1 score for POS tagging is in fact the tagging accuracy.", |
| "cite_spans": [ |
| { |
| "start": 107, |
| "end": 130, |
| "text": "(Zhang and Clark, 2008)", |
| "ref_id": "BIBREF39" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Implementation details", |
| "sec_num": "3.4" |
| }, |
| { |
| "text": "https://github.com/UKPLab/ emnlp2017-bilstm-cnn-crf 9 Pre-trained word and syllable embeddings are learned by training the Word2Vec Skip-gram model (Mikolov et al., 2013) Table 2 presents POS tagging accuracy and tagging speed of each model on the test set w.r.t. gold word segmentation, in which MarMoT is the most accurate model while RDRPOSTagger is the fastest one. In particular, MarMoT obtains 0.5%+ higher accuracy than the three BiLSTMbased models. This is not surprising as the training set of 27k sentences is relatively small compared to the training data available in other languages such as English or Chinese. Table 3 presents F1 scores for word segmentation and POS tagging in a real-world application scenario where the gold word-segmentation is not available. Comparing the results in Table 2 to results for the pipeline strategy, we observe a drop of about 2% for all models when using predicted segmentation instead of gold segmentation. Also, Table 3 clearly shows that the pipeline strategy helps produce better results than the joint strategy. In addition, pre-designed features in both RDR-POSTagger and MarMoT are designed to capture word-level information rather than syllable-level information, so it is also not surprising that for the joint strategy RDRPOSTagger is significantly lower while MarMoT is lower than the BiLSTM-CRF model with additional character-based representations. Tables 2 and 3 suggest that for a practical application to Vietnamese where performance accuracy is preferred, we should consider using the pipeline strategy with a traditional SOTA featurebased tagger such as MarMoT. If speed is preferred such as in big data, RDRPOSTagger would be a superior alternative. With the current state of training data available in Vietnamese, future research should focus on incorporating Vietnamese linguis- Table 3 : F1 scores (in %) for word segmentation (WSeg) and POS tagging (PTag) from unsegmented text. The pipeline strategy uses RDRsegmenter for word segmentation. In preliminary experiments, where we also train the five models above to predict a segmentation tag B or I for each syllable, we then find that RDRsegmenter obtains better word segmentation score than those five models. tic features into the traditional feature-based sequence taggers.", |
| "cite_spans": [ |
| { |
| "start": 148, |
| "end": 170, |
| "text": "(Mikolov et al., 2013)", |
| "ref_id": "BIBREF13" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 171, |
| "end": 178, |
| "text": "Table 2", |
| "ref_id": "TABREF2" |
| }, |
| { |
| "start": 624, |
| "end": 631, |
| "text": "Table 3", |
| "ref_id": null |
| }, |
| { |
| "start": 802, |
| "end": 809, |
| "text": "Table 2", |
| "ref_id": "TABREF2" |
| }, |
| { |
| "start": 963, |
| "end": 970, |
| "text": "Table 3", |
| "ref_id": null |
| }, |
| { |
| "start": 1411, |
| "end": 1425, |
| "text": "Tables 2 and 3", |
| "ref_id": "TABREF2" |
| }, |
| { |
| "start": 1849, |
| "end": 1856, |
| "text": "Table 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Implementation details", |
| "sec_num": "3.4" |
| }, |
| { |
| "text": "We have presented empirical comparisons between two strategies for Vietnamese POS tagging from unsegmented text and between SOTA feature-and neural network-based models. Experimental results on the benchmark Vietnamese treebank (Nguyen et al., 2009) show that the pipeline strategy produces higher scores of POS tagging from unsegmented text than the joint strategy. In addition, we also show that a traditional feature-based model (i.e. MarMoT) obtains better POS tagging accuracy than neural networkbased models. We provide a pre-trained MarMoT model for Vietnamese POS tagging at https:// github.com/datquocnguyen/VnMarMoT.", |
| "cite_spans": [ |
| { |
| "start": 228, |
| "end": 249, |
| "text": "(Nguyen et al., 2009)", |
| "ref_id": "BIBREF26" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "5" |
| }, |
| { |
| "text": "In the traditional underscore-based representation in Vietnamese word segmentation(Nguyen et al., 2009), white space is only used to separate words while underscore is used to separate syllables inside a word.Dat Quoc Nguyen, Thanh Vu, Dai Quoc Nguyen, Mark Dras and Mark Johnson. 2017. From Word Segmentation to POS Tagging for Vietnamese. In Proceedings of Australasian Language Technology Association Workshop, pages 108 113.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "The data was officially used for the Vietnamese POS tagging shared task at the second VLSP 2013 workshop.4 http://rdrpostagger.sourceforge.net5", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "https://github.com/XuezheMax/ LasagneNLP", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": "This research was partially supported by the Australian Government through the Australian Research Council's Discovery Projects funding scheme (project DP160102156). This research was also partially supported by NICTA, funded by the Australian Government through the Department of Communications and the Australian Research Council through the ICT Centre of Excellence Program.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acknowledgments", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Dual decomposition for Vietnamese part-of-speech tagging", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "Ngo Xuan", |
| "suffix": "" |
| }, |
| { |
| "first": "Kunihiko", |
| "middle": [], |
| "last": "Bach", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Hiraishi", |
| "suffix": "" |
| }, |
| { |
| "first": "Akira", |
| "middle": [], |
| "last": "Nguyen Le Minh", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Shimazu", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "Proceedings of the 17th International Conference on Knowledge Based and Intelligent Information and Engineering Systems", |
| "volume": "", |
| "issue": "", |
| "pages": "123--131", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ngo Xuan Bach, Kunihiko Hiraishi, Nguyen Le Minh, and Akira Shimazu. 2013. Dual decomposition for Vietnamese part-of-speech tagging. In Proceedings of the 17th International Conference on Knowledge Based and Intelligent Information and Engineering Systems. pages 123-131.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "A Philosophical Basis for Knowledge Acquisition", |
| "authors": [ |
| { |
| "first": "P", |
| "middle": [], |
| "last": "Compton", |
| "suffix": "" |
| }, |
| { |
| "first": "R", |
| "middle": [], |
| "last": "Jansen", |
| "suffix": "" |
| } |
| ], |
| "year": 1990, |
| "venue": "Knowledge Aquisition", |
| "volume": "2", |
| "issue": "3", |
| "pages": "241--257", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "P. Compton and R. Jansen. 1990. A Philosophical Ba- sis for Knowledge Acquisition. Knowledge Aquisi- tion 2(3):241-257.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "POS-Tagger for English-Vietnamese Bilingual Corpus", |
| "authors": [ |
| { |
| "first": "Dinh", |
| "middle": [], |
| "last": "Dien", |
| "suffix": "" |
| }, |
| { |
| "first": "Hoang", |
| "middle": [], |
| "last": "Kiem", |
| "suffix": "" |
| } |
| ], |
| "year": 2003, |
| "venue": "Proceedings of the HLT-NAACL 2003 Workshop on Building and Using Parallel Texts: Data Driven Machine Translation and Beyond", |
| "volume": "", |
| "issue": "", |
| "pages": "88--95", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dinh Dien and Hoang Kiem. 2003. POS-Tagger for English-Vietnamese Bilingual Corpus. In Proceed- ings of the HLT-NAACL 2003 Workshop on Build- ing and Using Parallel Texts: Data Driven Machine Translation and Beyond. pages 88-95.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Vietnamese Word Segmentation", |
| "authors": [ |
| { |
| "first": "Dinh", |
| "middle": [], |
| "last": "Dien", |
| "suffix": "" |
| }, |
| { |
| "first": "Hoang", |
| "middle": [], |
| "last": "Kiem", |
| "suffix": "" |
| }, |
| { |
| "first": "Nguyen", |
| "middle": [], |
| "last": "Van Toan", |
| "suffix": "" |
| } |
| ], |
| "year": 2001, |
| "venue": "Proceedings of the Sixth Natural Language Processing Pacific Rim Symposium", |
| "volume": "", |
| "issue": "", |
| "pages": "749--756", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dinh Dien, Hoang Kiem, and Nguyen Van Toan. 2001. Vietnamese Word Segmentation. In Proceedings of the Sixth Natural Language Processing Pacific Rim Symposium. pages 749-756.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "A Maximum Entropy Approach for Vietnamese Word Segmentation", |
| "authors": [ |
| { |
| "first": "Dien", |
| "middle": [], |
| "last": "Dinh", |
| "suffix": "" |
| }, |
| { |
| "first": "Thuy", |
| "middle": [], |
| "last": "Vu", |
| "suffix": "" |
| } |
| ], |
| "year": 2006, |
| "venue": "Proceedings of the 2006 International Conference on Research, Innovation and Vision for the Future", |
| "volume": "", |
| "issue": "", |
| "pages": "248--253", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dien Dinh and Thuy Vu. 2006. A Maximum Entropy Approach for Vietnamese Word Segmentation. In Proceedings of the 2006 International Conference on Research, Innovation and Vision for the Future. pages 248-253.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Incorporating Nesterov Momentum into Adam", |
| "authors": [ |
| { |
| "first": "Timothy", |
| "middle": [], |
| "last": "Dozat", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of the ICLR 2016 Workshop Track", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Timothy Dozat. 2016. Incorporating Nesterov Momen- tum into Adam. In Proceedings of the ICLR 2016 Workshop Track.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Bidirectional LSTM-CRF models for sequence tagging", |
| "authors": [ |
| { |
| "first": "Zhiheng", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| }, |
| { |
| "first": "Wei", |
| "middle": [], |
| "last": "Xu", |
| "suffix": "" |
| }, |
| { |
| "first": "Kai", |
| "middle": [], |
| "last": "Yu", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1508.01991" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Zhiheng Huang, Wei Xu, and Kai Yu. 2015. Bidi- rectional LSTM-CRF models for sequence tagging. arXiv preprint arXiv:1508.01991.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "An Error-Driven Word-Character Hybrid Model for Joint Chinese Word Segmentation and POS Tagging", |
| "authors": [ |
| { |
| "first": "Canasai", |
| "middle": [], |
| "last": "Kruengkrai", |
| "suffix": "" |
| }, |
| { |
| "first": "Kiyotaka", |
| "middle": [], |
| "last": "Uchimoto", |
| "suffix": "" |
| }, |
| { |
| "first": "Yiou", |
| "middle": [], |
| "last": "Jun'ichi Kazama", |
| "suffix": "" |
| }, |
| { |
| "first": "Kentaro", |
| "middle": [], |
| "last": "Wang", |
| "suffix": "" |
| }, |
| { |
| "first": "Hitoshi", |
| "middle": [], |
| "last": "Torisawa", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Isahara", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "Proceedings of the 47th Annual Meeting of the ACL and the 4th IJCNLP of the AFNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "513--521", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Canasai Kruengkrai, Kiyotaka Uchimoto, Jun'ichi Kazama, Yiou Wang, Kentaro Torisawa, and Hitoshi Isahara. 2009. An Error-Driven Word-Character Hybrid Model for Joint Chinese Word Segmentation and POS Tagging. In Proceedings of the 47th An- nual Meeting of the ACL and the 4th IJCNLP of the AFNLP. pages 513-521.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Neural Architectures for Named Entity Recognition", |
| "authors": [ |
| { |
| "first": "Guillaume", |
| "middle": [], |
| "last": "Lample", |
| "suffix": "" |
| }, |
| { |
| "first": "Miguel", |
| "middle": [], |
| "last": "Ballesteros", |
| "suffix": "" |
| }, |
| { |
| "first": "Sandeep", |
| "middle": [], |
| "last": "Subramanian", |
| "suffix": "" |
| }, |
| { |
| "first": "Kazuya", |
| "middle": [], |
| "last": "Kawakami", |
| "suffix": "" |
| }, |
| { |
| "first": "Chris", |
| "middle": [], |
| "last": "Dyer", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies", |
| "volume": "", |
| "issue": "", |
| "pages": "260--270", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Guillaume Lample, Miguel Ballesteros, Sandeep Sub- ramanian, Kazuya Kawakami, and Chris Dyer. 2016. Neural Architectures for Named Entity Recognition. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computa- tional Linguistics: Human Language Technologies. pages 260-270.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "A hybrid approach to word segmentation of Vietnamese texts", |
| "authors": [ |
| { |
| "first": "Thi", |
| "middle": [], |
| "last": "Hong Phuong Le", |
| "suffix": "" |
| }, |
| { |
| "first": "Azim", |
| "middle": [], |
| "last": "Minh Huyen Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Tuong Vinh", |
| "middle": [], |
| "last": "Roussanaly", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Ho", |
| "suffix": "" |
| } |
| ], |
| "year": 2008, |
| "venue": "Proceedings of the 2nd International Conference on Language and Automata Theory and Applications", |
| "volume": "", |
| "issue": "", |
| "pages": "240--249", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Hong Phuong Le, Thi Minh Huyen Nguyen, Azim Roussanaly, and Tuong Vinh Ho. 2008. A hybrid approach to word segmentation of Vietnamese texts. In Proceedings of the 2nd International Conference on Language and Automata Theory and Applica- tions. pages 240-249.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "An empirical study of maximum entropy approach for partof-speech tagging of Vietnamese texts", |
| "authors": [ |
| { |
| "first": "Phuong", |
| "middle": [], |
| "last": "Le-Hong", |
| "suffix": "" |
| }, |
| { |
| "first": "Azim", |
| "middle": [], |
| "last": "Roussanaly", |
| "suffix": "" |
| }, |
| { |
| "first": "Thi", |
| "middle": [], |
| "last": "Minh Huyen Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Mathias", |
| "middle": [], |
| "last": "Rossignol", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "Proceedings of the Traitement Automatique des Langues Naturelles", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Phuong Le-Hong, Azim Roussanaly, Thi Minh Huyen Nguyen, and Mathias Rossignol. 2010. An empir- ical study of maximum entropy approach for part- of-speech tagging of Vietnamese texts. In Proceed- ings of the Traitement Automatique des Langues Na- turelles.", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "Probabilistic Ensemble Learning for Vietnamese Word Segmentation", |
| "authors": [ |
| { |
| "first": "Wuying", |
| "middle": [], |
| "last": "Liu", |
| "suffix": "" |
| }, |
| { |
| "first": "Li", |
| "middle": [], |
| "last": "Lin", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "Proceedings of the 37th International ACM SIGIR Conference on Research & Development in Information Retrieval", |
| "volume": "", |
| "issue": "", |
| "pages": "931--934", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Wuying Liu and Li Lin. 2014. Probabilistic Ensemble Learning for Vietnamese Word Segmentation. In Proceedings of the 37th International ACM SIGIR Conference on Research & Development in Informa- tion Retrieval. pages 931-934.", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "End-to-end Sequence Labeling via Bi-directional LSTM-CNNs-CRF", |
| "authors": [ |
| { |
| "first": "Xuezhe", |
| "middle": [], |
| "last": "Ma", |
| "suffix": "" |
| }, |
| { |
| "first": "Eduard", |
| "middle": [], |
| "last": "Hovy", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "1064--1074", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Xuezhe Ma and Eduard Hovy. 2016. End-to-end Se- quence Labeling via Bi-directional LSTM-CNNs- CRF. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Vol- ume 1: Long Papers). pages 1064-1074.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "Distributed Representations of Words and Phrases and their Compositionality", |
| "authors": [ |
| { |
| "first": "Tomas", |
| "middle": [], |
| "last": "Mikolov", |
| "suffix": "" |
| }, |
| { |
| "first": "Ilya", |
| "middle": [], |
| "last": "Sutskever", |
| "suffix": "" |
| }, |
| { |
| "first": "Kai", |
| "middle": [], |
| "last": "Chen", |
| "suffix": "" |
| }, |
| { |
| "first": "Greg", |
| "middle": [ |
| "S" |
| ], |
| "last": "Corrado", |
| "suffix": "" |
| }, |
| { |
| "first": "Jeff", |
| "middle": [], |
| "last": "Dean", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "Advances in Neural Information Processing Systems 26", |
| "volume": "", |
| "issue": "", |
| "pages": "3111--3119", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S Cor- rado, and Jeff Dean. 2013. Distributed Representa- tions of Words and Phrases and their Composition- ality. In Advances in Neural Information Processing Systems 26. pages 3111-3119.", |
| "links": null |
| }, |
| "BIBREF14": { |
| "ref_id": "b14", |
| "title": "Efficient Higher-Order CRFs for Morphological Tagging", |
| "authors": [ |
| { |
| "first": "Thomas", |
| "middle": [], |
| "last": "Mueller", |
| "suffix": "" |
| }, |
| { |
| "first": "Helmut", |
| "middle": [], |
| "last": "Schmid", |
| "suffix": "" |
| }, |
| { |
| "first": "Hinrich", |
| "middle": [], |
| "last": "Sch\u00fctze", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "Proceedings of the 2013 Conference on Empirical Methods on Natural Language Processing", |
| "volume": "", |
| "issue": "", |
| "pages": "322--332", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Thomas Mueller, Helmut Schmid, and Hinrich Sch\u00fctze. 2013. Efficient Higher-Order CRFs for Morphological Tagging. In Proceedings of the 2013 Conference on Empirical Methods on Natural Language Processing. pages 322-332.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "A Chinese word segmentation model and a Chinese word segmentation system PC-CWSS", |
| "authors": [ |
| { |
| "first": "Liang", |
| "middle": [], |
| "last": "Nanyuan", |
| "suffix": "" |
| }, |
| { |
| "first": "Zheng", |
| "middle": [], |
| "last": "Yanbin", |
| "suffix": "" |
| } |
| ], |
| "year": 1991, |
| "venue": "Journal of Chinese Language and Computing", |
| "volume": "1", |
| "issue": "1", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Liang NanYuan and Zheng YanBin. 1991. A Chinese word segmentation model and a Chinese word seg- mentation system PC-CWSS. Journal of Chinese Language and Computing 1(1).", |
| "links": null |
| }, |
| "BIBREF16": { |
| "ref_id": "b16", |
| "title": "Improving Vietnamese POS tagging by integrating a rich feature set and Support Vector Machines", |
| "authors": [ |
| { |
| "first": "Minh", |
| "middle": [], |
| "last": "Nghiem", |
| "suffix": "" |
| }, |
| { |
| "first": "Dien", |
| "middle": [], |
| "last": "Dinh", |
| "suffix": "" |
| }, |
| { |
| "first": "Mai", |
| "middle": [], |
| "last": "Nguyen", |
| "suffix": "" |
| } |
| ], |
| "year": 2008, |
| "venue": "Proceedings of the 2008 IEEE International Conference on Research, Innovation and Vision for the Future in Computing and Communication Technologies", |
| "volume": "", |
| "issue": "", |
| "pages": "128--133", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Minh Nghiem, Dien Dinh, and Mai Nguyen. 2008. Im- proving Vietnamese POS tagging by integrating a rich feature set and Support Vector Machines. In Proceedings of the 2008 IEEE International Con- ference on Research, Innovation and Vision for the Future in Computing and Communication Technolo- gies. pages 128-133.", |
| "links": null |
| }, |
| "BIBREF17": { |
| "ref_id": "b17", |
| "title": "Vietnamese Word Segmentation with CRFs and SVMs: An Investigation", |
| "authors": [ |
| { |
| "first": "Cam-Tu", |
| "middle": [], |
| "last": "Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Trung-Kien", |
| "middle": [], |
| "last": "Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Xuan-Hieu", |
| "middle": [], |
| "last": "Phan", |
| "suffix": "" |
| }, |
| { |
| "first": "Le-Minh", |
| "middle": [], |
| "last": "Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Quang-Thuy", |
| "middle": [], |
| "last": "Ha", |
| "suffix": "" |
| } |
| ], |
| "year": 2006, |
| "venue": "Proceedings of the 20th", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Cam-Tu Nguyen, Trung-Kien Nguyen, Xuan-Hieu Phan, Le-Minh Nguyen, and Quang-Thuy Ha. 2006. Vietnamese Word Segmentation with CRFs and SVMs: An Investigation. In Proceedings of the 20th", |
| "links": null |
| }, |
| "BIBREF18": { |
| "ref_id": "b18", |
| "title": "Pacific Asia Conference on Language, Information and Computation", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "215--222", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Pacific Asia Conference on Language, Information and Computation. pages 215-222.", |
| "links": null |
| }, |
| "BIBREF19": { |
| "ref_id": "b19", |
| "title": "A Novel Neural Network Model for Joint POS Tagging and Graph-based Dependency Parsing", |
| "authors": [ |
| { |
| "first": "Mark", |
| "middle": [], |
| "last": "Dat Quoc Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Mark", |
| "middle": [], |
| "last": "Dras", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Johnson", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Proceedings of the CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies", |
| "volume": "", |
| "issue": "", |
| "pages": "134--142", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dat Quoc Nguyen, Mark Dras, and Mark Johnson. 2017a. A Novel Neural Network Model for Joint POS Tagging and Graph-based Dependency Pars- ing. In Proceedings of the CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Uni- versal Dependencies. pages 134-142.", |
| "links": null |
| }, |
| "BIBREF20": { |
| "ref_id": "b20", |
| "title": "RDRPOSTagger: A Ripple Down Rules-based Part-Of-Speech Tagger", |
| "authors": [], |
| "year": 2014, |
| "venue": "Proceedings of the Demonstrations at the 14th Conference of the European Chapter of the Association for Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "17--20", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dat Quoc Nguyen, Dai Quoc Nguyen, Dang Duc Pham, and Son Bao Pham. 2014a. RDRPOSTagger: A Rip- ple Down Rules-based Part-Of-Speech Tagger. In Proceedings of the Demonstrations at the 14th Con- ference of the European Chapter of the Association for Computational Linguistics. pages 17-20.", |
| "links": null |
| }, |
| "BIBREF21": { |
| "ref_id": "b21", |
| "title": "A Robust Transformation-Based Learning Approach Using Ripple Down Rules for Part-of-Speech Tagging", |
| "authors": [], |
| "year": 2016, |
| "venue": "AI Communications", |
| "volume": "29", |
| "issue": "3", |
| "pages": "409--422", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dat Quoc Nguyen, Dai Quoc Nguyen, Dang Duc Pham, and Son Bao Pham. 2016a. A Robust Transformation-Based Learning Approach Using Ripple Down Rules for Part-of-Speech Tagging. AI Communications 29(3):409-422.", |
| "links": null |
| }, |
| "BIBREF22": { |
| "ref_id": "b22", |
| "title": "From Treebank Conversion to Automatic Dependency Parsing for Vietnamese", |
| "authors": [], |
| "year": 2014, |
| "venue": "Proceedings of 19th International Conference on Application of Natural Language to Information Systems", |
| "volume": "", |
| "issue": "", |
| "pages": "196--207", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dat Quoc Nguyen, Dai Quoc Nguyen, Son Bao Pham, Phuong-Thai Nguyen, and Minh Le Nguyen. 2014b. From Treebank Conversion to Automatic Depen- dency Parsing for Vietnamese. In Proceedings of 19th International Conference on Application of Natural Language to Information Systems. pages 196-207.", |
| "links": null |
| }, |
| "BIBREF23": { |
| "ref_id": "b23", |
| "title": "Ripple Down Rules for Part-of-Speech Tagging", |
| "authors": [], |
| "year": 2011, |
| "venue": "Proceedings of the 12th International Conference on Intelligent Text Processing and Computational Linguistics -Volume Part I", |
| "volume": "", |
| "issue": "", |
| "pages": "190--201", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dat Quoc Nguyen, Dai Quoc Nguyen, Son Bao Pham, and Dang Duc Pham. 2011. Ripple Down Rules for Part-of-Speech Tagging. In Proceedings of the 12th International Conference on Intelligent Text Processing and Computational Linguistics -Volume Part I. pages 190-201.", |
| "links": null |
| }, |
| "BIBREF24": { |
| "ref_id": "b24", |
| "title": "2017b. A Fast and Accurate Vietnamese Word Segmenter", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1709.06307" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dat Quoc Nguyen, Dai Quoc Nguyen, Thanh Vu, Mark Dras, and Mark Johnson. 2017b. A Fast and Ac- curate Vietnamese Word Segmenter. arXiv preprint arXiv:1709.06307.", |
| "links": null |
| }, |
| "BIBREF25": { |
| "ref_id": "b25", |
| "title": "A Semi-supervised Learning Method for Vietnamese Part-of-Speech Tagging", |
| "authors": [ |
| { |
| "first": "Xuan", |
| "middle": [ |
| "Bach" |
| ], |
| "last": "Le Minh Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Ngo", |
| "suffix": "" |
| }, |
| { |
| "first": "Quang Nhat Minh", |
| "middle": [], |
| "last": "Viet Cuong Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Akira", |
| "middle": [], |
| "last": "Pham", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Shimazu", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "Proceedings of the International Conference on Knowledge and Systems Engineering", |
| "volume": "", |
| "issue": "", |
| "pages": "141--146", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Le Minh Nguyen, Xuan Bach Ngo, Viet Cuong Nguyen, Quang Nhat Minh Pham, and Akira Shi- mazu. 2010. A Semi-supervised Learning Method for Vietnamese Part-of-Speech Tagging. In Pro- ceedings of the International Conference on Knowl- edge and Systems Engineering. pages 141-146.", |
| "links": null |
| }, |
| "BIBREF26": { |
| "ref_id": "b26", |
| "title": "Building a Large Syntactically-Annotated Corpus of Vietnamese", |
| "authors": [ |
| { |
| "first": "Phuong", |
| "middle": [], |
| "last": "Thai Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Xuan", |
| "middle": [ |
| "Luong" |
| ], |
| "last": "Vu", |
| "suffix": "" |
| }, |
| { |
| "first": "Thi", |
| "middle": [], |
| "last": "Minh Huyen", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Hong", |
| "middle": [ |
| "Phuong" |
| ], |
| "last": "Van Hiep Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Le", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "Proceedings of the Third Linguistic Annotation Workshop", |
| "volume": "", |
| "issue": "", |
| "pages": "182--185", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Phuong Thai Nguyen, Xuan Luong Vu, Thi Minh Huyen Nguyen, Van Hiep Nguyen, and Hong Phuong Le. 2009. Building a Large Syntactically-Annotated Corpus of Vietnamese. In Proceedings of the Third Linguistic Annotation Workshop. pages 182-185.", |
| "links": null |
| }, |
| "BIBREF27": { |
| "ref_id": "b27", |
| "title": "A Hybrid Approach to Vietnamese Word Segmentation", |
| "authors": [ |
| { |
| "first": "Tuan-Phong", |
| "middle": [], |
| "last": "Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Anh-Cuong", |
| "middle": [], |
| "last": "Le", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of the 2016 IEEE RIVF International Conference on Computing and Communication Technologies: Research, Innovation, and Vision for the Future", |
| "volume": "", |
| "issue": "", |
| "pages": "114--119", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Tuan-Phong Nguyen and Anh-Cuong Le. 2016. A Hy- brid Approach to Vietnamese Word Segmentation. In Proceedings of the 2016 IEEE RIVF Interna- tional Conference on Computing and Communica- tion Technologies: Research, Innovation, and Vision for the Future. pages 114-119.", |
| "links": null |
| }, |
| "BIBREF28": { |
| "ref_id": "b28", |
| "title": "An Experimental Investigation of Part-Of-Speech Taggers for Vietnamese", |
| "authors": [ |
| { |
| "first": "Quoc", |
| "middle": [ |
| "Tuan" |
| ], |
| "last": "Tuan Phong Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Xuan", |
| "middle": [ |
| "Nam" |
| ], |
| "last": "Truong", |
| "suffix": "" |
| }, |
| { |
| "first": "Anh Cuong", |
| "middle": [], |
| "last": "Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Le", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "VNU Journal of Science: Computer Science and Communication Engineering", |
| "volume": "32", |
| "issue": "3", |
| "pages": "11--25", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Tuan Phong Nguyen, Quoc Tuan Truong, Xuan Nam Nguyen, and Anh Cuong Le. 2016b. An Experimen- tal Investigation of Part-Of-Speech Taggers for Viet- namese. VNU Journal of Science: Computer Sci- ence and Communication Engineering 32(3):11-25.", |
| "links": null |
| }, |
| "BIBREF29": { |
| "ref_id": "b29", |
| "title": "A Hybrid Approach to Vietnamese Word Segmentation using Part of Speech tags", |
| "authors": [ |
| { |
| "first": "Giang", |
| "middle": [], |
| "last": "Dang Duc Pham", |
| "suffix": "" |
| }, |
| { |
| "first": "Son", |
| "middle": [ |
| "Bao" |
| ], |
| "last": "Binh Tran", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Pham", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "Proceedings of the 2009 International Conference on Knowledge and Systems Engineering", |
| "volume": "", |
| "issue": "", |
| "pages": "154--161", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dang Duc Pham, Giang Binh Tran, and Son Bao Pham. 2009. A Hybrid Approach to Vietnamese Word Seg- mentation using Part of Speech tags. In Proceedings of the 2009 International Conference on Knowledge and Systems Engineering. pages 154-161.", |
| "links": null |
| }, |
| "BIBREF30": { |
| "ref_id": "b30", |
| "title": "NNVLP: A Neural Network-Based Vietnamese Language Processing Toolkit", |
| "authors": [ |
| { |
| "first": "Thai-Hoang", |
| "middle": [], |
| "last": "Pham", |
| "suffix": "" |
| }, |
| { |
| "first": "Xuan-Khoai", |
| "middle": [], |
| "last": "Pham", |
| "suffix": "" |
| }, |
| { |
| "first": "Tuan-Anh", |
| "middle": [], |
| "last": "Nguyen", |
| "suffix": "" |
| }, |
| { |
| "first": "Phuong", |
| "middle": [], |
| "last": "Le-Hong", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Proceedings of the IJCNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Thai-Hoang Pham, Xuan-Khoai Pham, Tuan-Anh Nguyen, and Phuong Le-Hong. 2017. NNVLP: A Neural Network-Based Vietnamese Language Pro- cessing Toolkit. In Proceedings of the IJCNLP 2017 System Demonstrations. page to appear.", |
| "links": null |
| }, |
| "BIBREF31": { |
| "ref_id": "b31", |
| "title": "Reporting Score Distributions Makes a Difference: Performance Study of LSTM-networks for Sequence Tagging", |
| "authors": [ |
| { |
| "first": "Nils", |
| "middle": [], |
| "last": "Reimers", |
| "suffix": "" |
| }, |
| { |
| "first": "Iryna", |
| "middle": [], |
| "last": "Gurevych", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing", |
| "volume": "", |
| "issue": "", |
| "pages": "338--348", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nils Reimers and Iryna Gurevych. 2017. Report- ing Score Distributions Makes a Difference: Perfor- mance Study of LSTM-networks for Sequence Tag- ging. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Process- ing. pages 338-348.", |
| "links": null |
| }, |
| "BIBREF32": { |
| "ref_id": "b32", |
| "title": "Character-based Joint Segmentation and POS Tagging for Chinese using Bidirectional RNN-CRF", |
| "authors": [ |
| { |
| "first": "Yan", |
| "middle": [], |
| "last": "Shao", |
| "suffix": "" |
| }, |
| { |
| "first": "Christian", |
| "middle": [], |
| "last": "Hardmeier", |
| "suffix": "" |
| }, |
| { |
| "first": "J\u00f6rg", |
| "middle": [], |
| "last": "Tiedemann", |
| "suffix": "" |
| }, |
| { |
| "first": "Joakim", |
| "middle": [], |
| "last": "Nivre", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Proceedings of the 8th International Joint Conference on Natural Language Processing", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Yan Shao, Christian Hardmeier, J\u00f6rg Tiedemann, and Joakim Nivre. 2017. Character-based Joint Segmen- tation and POS Tagging for Chinese using Bidirec- tional RNN-CRF. In Proceedings of the 8th Interna- tional Joint Conference on Natural Language Pro- cessing. page to appear.", |
| "links": null |
| }, |
| "BIBREF33": { |
| "ref_id": "b33", |
| "title": "Fundamental tools and resource are available for Vietnamese analysis", |
| "authors": [ |
| { |
| "first": "Kanji", |
| "middle": [], |
| "last": "Takahashi", |
| "suffix": "" |
| }, |
| { |
| "first": "Kazuhide", |
| "middle": [], |
| "last": "Yamamoto", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of the 2016 International Conference on Asian Language Processing", |
| "volume": "", |
| "issue": "", |
| "pages": "246--249", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kanji Takahashi and Kazuhide Yamamoto. 2016. Fun- damental tools and resource are available for Viet- namese analysis. In Proceedings of the 2016 Inter- national Conference on Asian Language Processing. pages 246-249.", |
| "links": null |
| }, |
| "BIBREF34": { |
| "ref_id": "b34", |
| "title": "Feature-rich Part-ofspeech Tagging with a Cyclic Dependency Network", |
| "authors": [ |
| { |
| "first": "Kristina", |
| "middle": [], |
| "last": "Toutanova", |
| "suffix": "" |
| }, |
| { |
| "first": "Dan", |
| "middle": [], |
| "last": "Klein", |
| "suffix": "" |
| }, |
| { |
| "first": "Christopher", |
| "middle": [ |
| "D" |
| ], |
| "last": "Manning", |
| "suffix": "" |
| }, |
| { |
| "first": "Yoram", |
| "middle": [], |
| "last": "Singer", |
| "suffix": "" |
| } |
| ], |
| "year": 2003, |
| "venue": "Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology", |
| "volume": "1", |
| "issue": "", |
| "pages": "173--180", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kristina Toutanova, Dan Klein, Christopher D. Man- ning, and Yoram Singer. 2003. Feature-rich Part-of- speech Tagging with a Cyclic Dependency Network. In Proceedings of the 2003 Conference of the North American Chapter of the Association for Computa- tional Linguistics on Human Language Technology - Volume 1. pages 173-180.", |
| "links": null |
| }, |
| "BIBREF35": { |
| "ref_id": "b35", |
| "title": "An effective context-based method for Vietnamese-word segmentation", |
| "authors": [ |
| { |
| "first": "Ngoc", |
| "middle": [ |
| "Anh" |
| ], |
| "last": "Tran", |
| "suffix": "" |
| }, |
| { |
| "first": "Thanh", |
| "middle": [ |
| "Tinh" |
| ], |
| "last": "Dao", |
| "suffix": "" |
| }, |
| { |
| "first": "Phuong", |
| "middle": [ |
| "Thai" |
| ], |
| "last": "Nguyen", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "Proceedings of First International Workshop on Vietnamese Language and Speech Processing", |
| "volume": "", |
| "issue": "", |
| "pages": "34--40", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ngoc Anh Tran, Thanh Tinh Dao, and Phuong Thai Nguyen. 2012. An effective context-based method for Vietnamese-word segmentation. In Proceedings of First International Workshop on Vietnamese Lan- guage and Speech Processing. pages 34-40.", |
| "links": null |
| }, |
| "BIBREF36": { |
| "ref_id": "b36", |
| "title": "An Experimental Study on Vietnamese POS Tagging", |
| "authors": [ |
| { |
| "first": "Cuong", |
| "middle": [ |
| "Anh" |
| ], |
| "last": "Oanh Thi Tran", |
| "suffix": "" |
| }, |
| { |
| "first": "Thuy", |
| "middle": [ |
| "Quang" |
| ], |
| "last": "Le", |
| "suffix": "" |
| }, |
| { |
| "first": "Quynh Hoang", |
| "middle": [], |
| "last": "Ha", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Le", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "Proceedings of the 2009 International Conference on Asian Language Processing", |
| "volume": "", |
| "issue": "", |
| "pages": "23--27", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Oanh Thi Tran, Cuong Anh Le, Thuy Quang Ha, and Quynh Hoang Le. 2009. An Experimental Study on Vietnamese POS Tagging. In Proceedings of the 2009 International Conference on Asian Language Processing. pages 23-27.", |
| "links": null |
| }, |
| "BIBREF37": { |
| "ref_id": "b37", |
| "title": "Improving Vietnamese Word Segmentation and POS Tagging using MEM with Various Kinds of Resources", |
| "authors": [ |
| { |
| "first": "Thi", |
| "middle": [], |
| "last": "Oanh Tran", |
| "suffix": "" |
| }, |
| { |
| "first": "Anh", |
| "middle": [ |
| "Cuong" |
| ], |
| "last": "Le", |
| "suffix": "" |
| }, |
| { |
| "first": "Quang Thuy", |
| "middle": [], |
| "last": "Ha", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "Journal of Natural Language Processing", |
| "volume": "17", |
| "issue": "3", |
| "pages": "41--60", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Thi Oanh Tran, Anh Cuong Le, and Quang Thuy Ha. 2010. Improving Vietnamese Word Segmentation and POS Tagging using MEM with Various Kinds of Resources. Journal of Natural Language Processing 17(3):41-60.", |
| "links": null |
| }, |
| "BIBREF38": { |
| "ref_id": "b38", |
| "title": "Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies", |
| "authors": [ |
| { |
| "first": "Daniel", |
| "middle": [], |
| "last": "Zeman", |
| "suffix": "" |
| }, |
| { |
| "first": "Martin", |
| "middle": [], |
| "last": "Popel", |
| "suffix": "" |
| }, |
| { |
| "first": "Milan", |
| "middle": [], |
| "last": "Straka", |
| "suffix": "" |
| }, |
| { |
| "first": "Jan", |
| "middle": [], |
| "last": "Hajic", |
| "suffix": "" |
| }, |
| { |
| "first": "Joakim", |
| "middle": [], |
| "last": "Nivre", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Proceedings of the CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies", |
| "volume": "", |
| "issue": "", |
| "pages": "1--19", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Daniel Zeman, Martin Popel, Milan Straka, Jan Ha- jic, Joakim Nivre, et al. 2017. CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Uni- versal Dependencies. In Proceedings of the CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies. pages 1-19.", |
| "links": null |
| }, |
| "BIBREF39": { |
| "ref_id": "b39", |
| "title": "Joint Word Segmentation and POS Tagging Using a Single Perceptron", |
| "authors": [ |
| { |
| "first": "Yue", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| }, |
| { |
| "first": "Stephen", |
| "middle": [], |
| "last": "Clark", |
| "suffix": "" |
| } |
| ], |
| "year": 2008, |
| "venue": "Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies", |
| "volume": "", |
| "issue": "", |
| "pages": "888--896", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Yue Zhang and Stephen Clark. 2008. Joint Word Seg- mentation and POS Tagging Using a Single Percep- tron. In Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics: Hu- man Language Technologies. pages 888-896.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "TABREF1": { |
| "type_str": "table", |
| "num": null, |
| "text": "on a Vietnamese news corpus which is available at: http://mim.hus.vnu.edu.vn/phuonglh/ corpus/baomoi.zip", |
| "html": null, |
| "content": "<table><tr><td>Model RDRPOSTagger MarMoT</td><td colspan=\"2\">Accuracy Speed 95.11 180k 95.88 25k</td></tr><tr><td>BiLSTM-CRF</td><td>95.06</td><td>3k</td></tr><tr><td>+ CNN-char</td><td>95.40</td><td>2.5k</td></tr><tr><td>+ LSTM-char</td><td>95.31</td><td>1.5k</td></tr></table>" |
| }, |
| "TABREF2": { |
| "type_str": "table", |
| "num": null, |
| "text": "", |
| "html": null, |
| "content": "<table><tr><td>: POS tagging accuracies (in %) on the test</td></tr><tr><td>set w.r.t. gold word segmentation. \"Speed\" denotes</td></tr><tr><td>the tagging speed, i.e. the number of words per</td></tr><tr><td>second, computed on a personal computer of Intel</td></tr><tr><td>Core i7 2.2 GHz (model loading time is not taken</td></tr><tr><td>into account).</td></tr></table>" |
| } |
| } |
| } |
| } |
