---
title: "Joint Word Segmentation and POS Tagging Using a Single"
description: "Zhang and Clark build a single discriminative linear model over full segmented-and-tagged outputs. 14 references, cited by 64 papers here."
authors: ["Yue Zhang", "Stephen Clark"]
year: 2008
venue: "Proceedings of ACL-08: HLT"
source: "https://aclanthology.org/P08-1101/"
licence: "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)"
licence_url: "https://creativecommons.org/licenses/by-nc-sa/3.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/zhang-2008-joint-word-segmentation-pos-tagging/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zhang-2008-joint-word-segmentation-pos-tagging/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2008](https://md.research.elamur.ai/free/en/papers/2008/) / Joint Word Segmentation and POS Tagging Using a Single Perceptron

# Joint Word Segmentation and POS Tagging Using a Single Perceptron

**Authors:** [Yue Zhang](https://md.research.elamur.ai/free/en/authors/yue-zhang/) and [Stephen Clark](https://md.research.elamur.ai/free/en/authors/stephen-clark/) · **Year:** 2008 · **Venue:** [Proceedings of ACL-08: HLT](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2008](https://md.research.elamur.ai/free/en/venues/acl/2008/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY-NC-SA 3.0 (ACL Anthology, before 2016)](https://creativecommons.org/licenses/by-nc-sa/3.0/)

**Links:** [Publisher's page](https://aclanthology.org/P08-1101/) · [PDF](https://aclanthology.org/P08-1101.pdf) · [OpenAlex](https://openalex.org/W2160097208)

**Citations:** 123 citations (OpenAlex, 2026-10-14) · [64](https://md.research.elamur.ai/free/en/papers/zhang-2008-joint-word-segmentation-pos-tagging/cited-by/) in our library, of the 141,840 papers whose bibliographies we read

## Summary
Chinese text has no explicit word boundaries, so POS tagging usually runs after a separate segmentation step. Segmentation mistakes then pass into the tagger and cannot be fixed. Earlier joint systems shrank the combined search space by tagging at the character level (Ng and Low, 2004) or by reranking N-best segmentations (Shi and Wang, 2007). Both approaches limit how much word-level and POS evidence can interact. Zhang and Clark build a single discriminative linear model over full segmented-and-tagged outputs. Its feature set is simply the union of the templates from their baseline perceptron segmentor and baseline POS tagger. The model is trained with one averaged perceptron. Because a standard beam search decoder performs poorly on this joint space, they introduce a multiple-beam search algorithm. On the Chinese Treebank (CTB 3), evaluated with 10-fold cross validation, the joint model cuts segmentation errors by about 14.6% and overall tagging errors by about 12.2% relative to the pipeline. Its accuracy is competitive with systems that rely on outside resources, even though it is purely data-driven.

## Key points
- With the 10-fold split, the pipeline baseline reaches 95.20 segmentation F-score (SF), 90.33 overall F-score (TF) and 92.17 tagging accuracy by character (TA). The joint model reaches 95.90 SF, 91.34 TF and 93.02 TA, and it beats the baseline in every one of the ten folds.
- The single-beam decoder from Zhang and Clark (2007) was enough for segmentation alone with an agenda of 8. In the joint model it lowered segmentation accuracy below the baseline segmentor, even with a beam size of 1024.
- Shi and Wang (2007) report slightly higher absolute scores (96.05 SF, 91.86 TF). However, they use special-character knowledge and a semantic net from outside CTB, and their data partition differs. The authors follow Ng and Low's even split of sentences ordered by sentence ID.
- In the error analysis, confusions between NN and VV (in both directions) and between NR and NN make up 14.92% of all errors. Roughly half of all errors (51.47%) are segmentation errors.
- The authors also tried SVM struct for training the parameters, but it was far too slow to be practical.

## Interesting ideas
- **Fair joint-vs-pipeline comparison:** the joint model uses only the features already present in the two pipeline components. Any gain can therefore be credited to joint inference rather than to extra information. This is a clean experimental design for any study of whether merging two tasks helps.
- **Rethinking which hypotheses compete in a beam:** the authors asked which partial candidates are fairly comparable when items mix words and tags. Their answer was to keep several agendas instead of a single beam ranked across everything. This lesson carries over to other joint decoding problems where hypotheses differ in structure.
- **Pruning the search with training statistics:** the decoder limits where a word may start using a maximum observed word length for each tag. A tag dictionary further restricts frequent words and closed-set tags to the tags seen in training. Both are cheap constraints that speed up decoding without hand-written rules.
- **Character categories as features:** a character's category is defined as the set of tags it carried in the training data. This gives morphology-like features without needing external web corpora, which is attractive for languages with few resources.
- **POS features that also inform segmentation:** in a joint model, a tagger feature such as a number word followed by a measure word also signals where word boundaries fall. Features from one task can thus supervise the other at no extra cost.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [University of Oxford](https://md.research.elamur.ai/free/en/universities/university-of-oxford/) — University, Oxford, United Kingdom · 52 works · 154,829 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,086 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,889 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 9,296 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 9,206 works

🧭 Topics (2): [🔤 Part-of-speech tagging](https://md.research.elamur.ai/free/en/topics/part-of-speech-tagging/) · 91 works · [✂️ Tokenization & segmentation](https://md.research.elamur.ai/free/en/topics/tokenization/) · 77 works

## Cite

Cite as: Yue Zhang and Stephen Clark. 2008. Joint Word Segmentation and POS Tagging Using a Single Perceptron. Proceedings of ACL-08: HLT.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{zhang2008joint,
  title = {{Joint Word Segmentation and POS Tagging Using a Single Perceptron}},
  author = {Yue Zhang and Stephen Clark},
  booktitle = {Proceedings of ACL-08: HLT},
  year = {2008},
  url = {https://aclanthology.org/P08-1101/}
}
```

- [← previous Learning Morphology with Morfette](https://md.research.elamur.ai/free/en/papers/chrupala-2008-learning-morphology-morfette/)
- [next → Applying Morphology Generation Models to Machine Translation](https://md.research.elamur.ai/free/en/papers/toutanova-2008-applying-morphology-generation-models-machine/)
- [↑ back to papers of 2008](https://md.research.elamur.ai/free/en/papers/2008/)

## References

14 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Michael Collins. 2002. [Discriminative training methods for hidden Markov models: Theory and experiments with perceptron algorithms](https://md.research.elamur.ai/free/en/papers/collins-2002-discriminative-training-methods-hidden-markov/). In Proceedings of the EMNLP conference, pages 1–8, Philadelphia, PA.
2. Hal Daumé III and Daniel Marcu. 2005. Learning as search optimization: Approximate large margin methods for structured prediction. In Proceedings of the ICML Conference, pages 169–176, Bonn, Germany. [find](https://scholar.google.com/scholar?q=Hal+Daume%CC%81+III+and+Daniel+Marcu%2E+2005%2E+Learning+as+search+optimization%3A+Approximate+large+margin+methods+for+structured+prediction%2E+In+Proceedings+of+the+ICML+Conference%2C+pages+169%E2%80%93176%2C+Bonn%2C+Germany%2E)
3. Jenny Rose Finkel, Christopher D. Manning, and Andrew Y. Ng. 2006. Solving the problem of cascading errors: Approximate Bayesian inference for linguistic annotation pipelines. In Proceedings of the EMNLP Conference, pages 618–626, Sydney, Australia. [find](https://scholar.google.com/scholar?q=Jenny+Rose+Finkel%2C+Christopher+D%2E+Manning%2C+and+Andrew+Y%2E+Ng%2E+2006%2E+Solving+the+problem+of+cascading+errors%3A+Approximate+Bayesian+inference+for+linguistic+annotation+pipelines%2E+In+Proceedings+of+the+EM)
4. Tetsuji Nakagawa and Kiyotaka Uchimoto. 2007. A hybrid approach to word segmentation and pos tagging. In Proceedings of ACL Demo and Poster Session, pages 217–220, Prague, Czech Republic. [find](https://scholar.google.com/scholar?q=Tetsuji+Nakagawa+and+Kiyotaka+Uchimoto%2E+2007%2E+A+hybrid+approach+to+word+segmentation+and+pos+tagging%2E+In+Proceedings+of+ACL+Demo+and+Poster+Session%2C+pages+217%E2%80%93220%2C+Prague%2C+Czech+Republic%2E)
5. Hwee Tou Ng and Jin Kiat Low. 2004. [Chinese part-of-speech tagging: One-at-a-time or all-at-once? Word-based or character-based?](https://md.research.elamur.ai/free/en/papers/ng-2004-chinese-part-speech-tagging-one/) In Proceedings of the EMNLP Conference, pages 277–284, Barcelona, Spain.
6. Adwait Ratnaparkhi. 1996. [A maximum entropy model for part-of-speech tagging](https://md.research.elamur.ai/free/en/papers/ratnaparkhi-1996-maximum-entropy-model-part-speech/). In Proceedings of the EMNLP Conference, pages 133–142, Philadelphia, PA.
7. Murat Saraclar and Brian Roark. 2005. Joint discriminative language modeling and utterance classification. In Proceedings of the ICASSP Conference, volume 1, Philadelphia, USA. [find](https://scholar.google.com/scholar?q=Murat+Saraclar+and+Brian+Roark%2E+2005%2E+Joint+discriminative+language+modeling+and+utterance+classification%2E+In+Proceedings+of+the+ICASSP+Conference%2C+volume+1%2C+Philadelphia%2C+USA%2E)
8. Yanxin Shi and Mengqiu Wang. 2007. A dual-layer CRF based joint decoding method for cascade segmentation and labelling tasks. In Proceedings of the IJCAI Conference, Hyderabad, India. [find](https://scholar.google.com/scholar?q=Yanxin+Shi+and+Mengqiu+Wang%2E+2007%2E+A+dual%2Dlayer+CRF+based+joint+decoding+method+for+cascade+segmentation+and+labelling+tasks%2E+In+Proceedings+of+the+IJCAI+Conference%2C+Hyderabad%2C+India%2E)
9. Nobuyuki Shimizu and Andrew Haas. 2006. Exact decoding for jointly labeling and chunking sequences. In Proceedings of the COLING/ACL Conference, Poster Sessions, Sydney, Australia. [find](https://scholar.google.com/scholar?q=Nobuyuki+Shimizu+and+Andrew+Haas%2E+2006%2E+Exact+decoding+for+jointly+labeling+and+chunking+sequences%2E+In+Proceedings+of+the+COLING%2FACL+Conference%2C+Poster+Sessions%2C+Sydney%2C+Australia%2E)
10. Charles Sutton, Khashayar Rohanimanesh, and Andrew McCallum. 2004. Dynamic conditional random fields: Factorized probabilistic models for labeling and segmenting sequence data. In Proceedings of the ICML Conference, Banff, Canada. [find](https://scholar.google.com/scholar?q=Charles+Sutton%2C+Khashayar+Rohanimanesh%2C+and+Andrew+McCallum%2E+2004%2E+Dynamic+conditional+random+fields%3A+Factorized+probabilistic+models+for+labeling+and+segmenting+sequence+data%2E+In+Proceedings+of+the+I)
11. Huihsin Tseng, Daniel Jurafsky, and Christopher Manning. 2005. Morphological features help POS tagging of unknown words across language varieties. In Proceedings of the Fourth SIGHAN Workshop, Jeju Island, Korea. [find](https://scholar.google.com/scholar?q=Huihsin+Tseng%2C+Daniel+Jurafsky%2C+and+Christopher+Manning%2E+2005%2E+Morphological+features+help+POS+tagging+of+unknown+words+across+language+varieties%2E+In+Proceedings+of+the+Fourth+SIGHAN+Workshop%2C+Jeju+Is)
12. I. Tsochantaridis, T. Hofmann, T. Joachims, and Y. Altun. 2004. Support vector machine learning for interdependent and structured output spaces. In Proceedings of the ICML Conference, Banff, Canada. [find](https://scholar.google.com/scholar?q=I%2E+Tsochantaridis%2C+T%2E+Hofmann%2C+T%2E+Joachims%2C+and+Y%2E+Altun%2E+2004%2E+Support+vector+machine+learning+for+interdependent+and+structured+output+spaces%2E+In+Proceedings+of+the+ICML+Conference%2C+Banff%2C+Canada%2E)
13. Fei Xia. 2000. The part-of-speech tagging guidelines for the Chinese Treebank (3.0). IRCS Report, University of Pennsylvania. [find](https://scholar.google.com/scholar?q=Fei+Xia%2E+2000%2E+The+part%2Dof%2Dspeech+tagging+guidelines+for+the+Chinese+Treebank+%283%2E0%29%2E+IRCS+Report%2C+University+of+Pennsylvania%2E)
14. Yue Zhang and Stephen Clark. 2007. Chinese segmentation with a word-based perceptron algorithm. In Proceedings of the ACL Conference, pages 840–847, Prague, Czech Republic. [find](https://scholar.google.com/scholar?q=Yue+Zhang+and+Stephen+Clark%2E+2007%2E+Chinese+segmentation+with+a+word%2Dbased+perceptron+algorithm%2E+In+Proceedings+of+the+ACL+Conference%2C+pages+840%E2%80%93847%2C+Prague%2C+Czech+Republic%2E)

*Updated 24 October 2026*