---
title: "Unsupervised Semantic Parsing — Poon 2009"
description: "Semantic parsing turns sentences into formal logical meaning representations."
authors: ["Hoifung Poon", "Pedro Domingos"]
year: 2009
venue: "Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing"
source: "https://aclanthology.org/D09-1001/"
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/poon-2009-unsupervised-semantic-parsing/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/poon-2009-unsupervised-semantic-parsing/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2009](https://md.research.elamur.ai/free/en/papers/2009/) / Unsupervised Semantic Parsing

# Unsupervised Semantic Parsing

**Authors:** [Hoifung Poon](https://md.research.elamur.ai/free/en/authors/hoifung-poon/), [Pedro Domingos](https://md.research.elamur.ai/free/en/authors/pedro-domingos/) · **Year:** 2009 · **Venue:** Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing · **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/D09-1001/) · [PDF](https://aclanthology.org/D09-1001.pdf) · [OpenAlex](https://openalex.org/W2150406842)

**Citations:** 253 citations (openalex-snapshot · 2026-10-14) · [92](#cited-by) in our library, of the 140884 papers whose bibliographies we read

## Summary
Semantic parsing turns sentences into formal logical meaning representations. Earlier learned parsers were supervised: each training sentence needed a target logical form, which is costly, hard to keep consistent, and nearly impossible for open-domain text. Poon and Domingos present USP, described as the first unsupervised semantic parser, built on Markov logic. USP does not learn syntax and semantics jointly. It takes dependency trees as input and deterministically rewrites them into quasi-logical forms (QLFs): each word becomes a unary atom (lemma plus POS tag), and each dependency edge becomes a binary atom. Sub-formulas become lambda forms, which are clustered recursively. Clustering starts from atoms of the same token type and grows to larger expressions whose parts share clusters, so paraphrases such as active and passive voice end up with one logical form. Parsing finds the MAP assignment of sentence parts to clusters and composes them. The authors built a knowledge base from GENIA biomedical abstracts and answered questions over it. USP produced 295 correct answers at 88% accuracy, clearly beating TextRunner, RESOLVER, DIRT and keyword baselines. This shows that deep meaning can be induced without any labels.

## Key points
- USP uses a Markov logic network. Its learned objects are lambda-form clusters, which stand in for the predicate and object constants that supervised parsing would fix in advance.
- The input is dependency trees. Nodes become atoms like `Utah(n2)`, edges become atoms like `nsubj(n1, n2)`, and the QLF of a sentence is the conjunction of all these atoms, with Skolem constants per node.
- Evaluation used question answering over a knowledge base extracted from the GENIA biomedical corpus. Question parses are matched to knowledge-base parses by subsumption.
- On Table 1, USP returned 334 answers with 295 correct (88%). The next most productive system, RS-SUB, returned 196 answers at 41%. DIRT had the second most correct answers (94, 59%), and KW-SYN reached 77% with 67 correct.
- Learned clusters include groups such as {diminish, reduce, decrease, attenuate} and {synthesis, production, secretion, release}. USP also equates argument variants like "expression of X" and "X expression".

## Interesting ideas
- **Predicates as clusters of paraphrases.** USP does not predefine a relation inventory. It treats each predicate or entity as a learned set of interchangeable surface forms. This suits any domain where no ontology exists in advance, such as trade or geographic texts.
- **Separate syntax from semantics.** USP builds meaning on top of an existing dependency parser instead of learning both at once. This lets semantic work benefit from improvements in parsers. For example, the same pipeline could start from UD-style parses for a language with a decent treebank.
- **Cheap lambda abstraction.** A constant that is not introduced by a unary atom inside a fragment is turned into an argument variable. This mechanical rule yields argument slots without hand-written lexical entries, and could be reused for extracting relation templates from trees.
- **Recursive relational clustering.** Forms are grouped not only by shared arguments but also by being built from sub-forms in the same clusters. Similarity then propagates bottom-up through structure, which is a useful pattern for paraphrase or relation mining.
- **Extrinsic QA evaluation with an informed baseline.** Without gold logical forms, the authors judge parses by how many questions they answer correctly. They also include a syntax-aware keyword baseline (KW-SYN), which turned out to be surprisingly strong. Both choices are a sensible way to evaluate unsupervised meaning extraction honestly.

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

Fields: [Natural language processing](https://md.research.elamur.ai/free/en/shelves/nlp/)

Topics: [Dependency parsing](https://md.research.elamur.ai/free/en/topics/dependency-parsing/)

## Cited by

Cited by

92 papers of our library cite it (of the 140884 whose bibliographies we read); 5 of them have a page here, 87 not yet.

- [Learning to Solve Arithmetic Word Problems with Verb Categorization](https://md.research.elamur.ai/free/en/papers/hosseini-2014-learning-solve-arithmetic-word-problems/)
- [One Vector is Not Enough: Entity-Augmented Distributed Semantics for Discourse Relations](https://md.research.elamur.ai/free/en/papers/ji-2015-one-vector-is-not-enough/)
- [Lexical Generalization in CCG Grammar Induction for Semantic Parsing](https://md.research.elamur.ai/free/en/papers/kwiatkowski-2011-lexical-generalization-ccg-grammar-induction/)
- [Annotated Gigaword](https://md.research.elamur.ai/free/en/papers/napoles-2012-annotated-gigaword/)
- [Enhanced English Universal Dependencies: An Improved Representation for Natural Language Understanding Tasks](https://md.research.elamur.ai/free/en/papers/ud-2016-schuster-enhanced-english-universal-dependencies-improved/)

## Cite

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

```bibtex
@inproceedings{poon2009unsupervised,
  title = {{Unsupervised Semantic Parsing}},
  author = {Poon, Hoifung and Domingos, Pedro},
  booktitle = {Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing},
  year = {2009},
  doi = {10.3115/1699510.1699512},
  url = {https://doi.org/10.3115/1699510.1699512}
}
```

- [← previous Unsupervised Recognition of Literal and Non-Literal Use of Idiomatic Expressions](https://md.research.elamur.ai/free/en/papers/sporleder-2009-unsupervised-recognition-literal-non-literal/)
- [next → Unsupervised Type and Token Identification of Idiomatic Expressions](https://md.research.elamur.ai/free/en/papers/fazly-2009-unsupervised-type-token-identification-idiomatic/)
- [↑ back to papers of 2009](https://md.research.elamur.ai/free/en/papers/2009/)

## References

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

1. G. Bakir, T. Hofmann, B. B. Schölkopf, A. Smola, B. Taskar, S. Vishwanathan, and (eds.). 2007. Predicting Structured Data. MIT Press, Cambridge, MA. [find](https://scholar.google.com/scholar?q=G%2E+Bakir%2C+T%2E+Hofmann%2C+B%2E+B%2E+Scho%CC%88lkopf%2C+A%2E+Smola%2C+B%2E+Taskar%2C+S%2E+Vishwanathan%2C+and+%28eds%2E%29%2E+2007%2E+Predicting+Structured+Data%2E+MIT+Press%2C+Cambridge%2C+MA%2E)
2. Michele Banko, Michael J. Cafarella, Stephen Soderland, Matt Broadhead, and Oren Etzioni. 2007. Open information extraction from the web. In Proceedings of the Twentieth International Joint Conference on Artificial Intelligence, pages 2670–2676, Hyderabad, India. AAAI Press. [find](https://scholar.google.com/scholar?q=Michele+Banko%2C+Michael+J%2E+Cafarella%2C+Stephen+Soderland%2C+Matt+Broadhead%2C+and+Oren+Etzioni%2E+2007%2E+Open+information+extraction+from+the+web%2E+In+Proceedings+of+the+Twentieth+International+Joint+Conference) [our card](https://md.research.elamur.ai/free/en/papers/banko-2007-open-information-extraction-web/)
3. Xavier Carreras and Luis Marquez. 2004. Introduction to the CoNLL-2004 shared task: Semantic role labeling. In Proceedings of the Eighth Conference on Computational Natural Language Learning, pages 89–97, Boston, MA. ACL. [find](https://scholar.google.com/scholar?q=Xavier+Carreras+and+Luis+Marquez%2E+2004%2E+Introduction+to+the+CoNLL%2D2004+shared+task%3A+Semantic+role+labeling%2E+In+Proceedings+of+the+Eighth+Conference+on+Computational+Natural+Language+Learning%2C+pages+89)
4. Marie-Catherine de Marneffe, Bill MacCartney, and Christopher D. Manning. 2006. Generating typed dependency parses from phrase structure parses. In Proceedings of the Fifth International Conference on Language Resources and Evaluation, pages 449– 454, Genoa, Italy. ELRA. [find](https://scholar.google.com/scholar?q=Marie%2DCatherine+de+Marneffe%2C+Bill+MacCartney%2C+and+Christopher+D%2E+Manning%2E+2006%2E+Generating+typed+dependency+parses+from+phrase+structure+parses%2E+In+Proceedings+of+the+Fifth+International+Conference+on) [our card](https://md.research.elamur.ai/free/en/papers/map-2006-marneffe-generating-typed-dependency-parses-phrase/)
5. Ruifang Ge and Raymond J. Mooney. 2009. Learning a compositional semantic parser using an existing syntactic parser. In Proceedings of the Forty Seventh Annual Meeting of the Association for Computational Linguistics, Singapore. ACL. [find](https://scholar.google.com/scholar?q=Ruifang+Ge+and+Raymond+J%2E+Mooney%2E+2009%2E+Learning+a+compositional+semantic+parser+using+an+existing+syntactic+parser%2E+In+Proceedings+of+the+Forty+Seventh+Annual+Meeting+of+the+Association+for+Computati)
6. Lise Getoor and Ben Taskar, editors. 2007. Introduction to Statistical Relational Learning. MIT Press, Cambridge, MA. [find](https://scholar.google.com/scholar?q=Lise+Getoor+and+Ben+Taskar%2C+editors%2E+2007%2E+Introduction+to+Statistical+Relational+Learning%2E+MIT+Press%2C+Cambridge%2C+MA%2E)
7. Pekka Kilpelainen. 1992. Tree Matching Problems with Applications to Structured Text databases. Ph.D. Thesis, Department of Computer Science, University of Helsinki. [find](https://scholar.google.com/scholar?q=Pekka+Kilpelainen%2E+1992%2E+Tree+Matching+Problems+with+Applications+to+Structured+Text+databases%2E+Ph%2ED%2E+Thesis%2C+Department+of+Computer+Science%2C+University+of+Helsinki%2E)
8. Jin-Dong Kim, Tomoko Ohta, Yuka Tateisi, and Jun’ichi Tsujii. 2003. GENIA corpus - a semantically annotated corpus for bio-textmining. Bioinformatics, 19:180–82. [find](https://scholar.google.com/scholar?q=Jin%2DDong+Kim%2C+Tomoko+Ohta%2C+Yuka+Tateisi%2C+and+Jun%E2%80%99ichi+Tsujii%2E+2003%2E+GENIA+corpus+%2D+a+semantically+annotated+corpus+for+bio%2Dtextmining%2E+Bioinformatics%2C+19%3A180%E2%80%9382%2E)
9. Stanley Kok and Pedro Domingos. 2008. Extracting semantic networks from text via relational clustering. In Proceedings of the Nineteenth European Conference on Machine Learning, pages 624–639, Antwerp, Belgium. Springer. [find](https://scholar.google.com/scholar?q=Stanley+Kok+and+Pedro+Domingos%2E+2008%2E+Extracting+semantic+networks+from+text+via+relational+clustering%2E+In+Proceedings+of+the+Nineteenth+European+Conference+on+Machine+Learning%2C+pages+624%E2%80%93639%2C+Antwerp)
10. Percy Liang and Dan Klein. 2008. Analyzing the errors of unsupervised learning. In Proceedings of the Forty Sixth Annual Meeting of the Association for Computational Linguistics, pages 879–887, Columbus, OH. ACL. [find](https://scholar.google.com/scholar?q=Percy+Liang+and+Dan+Klein%2E+2008%2E+Analyzing+the+errors+of+unsupervised+learning%2E+In+Proceedings+of+the+Forty+Sixth+Annual+Meeting+of+the+Association+for+Computational+Linguistics%2C+pages+879%E2%80%93887%2C+Columb)
11. Dekang Lin and Patrick Pantel. 2001. DIRT - discovery of inference rules from text. In Proceedings of the Seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 323–328, San Francisco, CA. ACM Press. [find](https://scholar.google.com/scholar?q=Dekang+Lin+and+Patrick+Pantel%2E+2001%2E+DIRT+%2D+discovery+of+inference+rules+from+text%2E+In+Proceedings+of+the+Seventh+ACM+SIGKDD+International+Conference+on+Knowledge+Discovery+and+Data+Mining%2C+pages+323%E2%80%93)
12. Dekang Lin. 1998. Dependency-based evaluation of MINIPAR. In Proceedings of the Workshop on the Evaluation of Parsing Systems, Granada, Spain. ELRA. [find](https://scholar.google.com/scholar?q=Dekang+Lin%2E+1998%2E+Dependency%2Dbased+evaluation+of+MINIPAR%2E+In+Proceedings+of+the+Workshop+on+the+Evaluation+of+Parsing+Systems%2C+Granada%2C+Spain%2E+ELRA%2E)
13. Saif Mohammad, Bonnie Dorr, and Graeme Hirst. 2008. Computing word-pair antonymy. In Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing, pages 982–991, Honolulu, HI. ACL. [find](https://scholar.google.com/scholar?q=Saif+Mohammad%2C+Bonnie+Dorr%2C+and+Graeme+Hirst%2E+2008%2E+Computing+word%2Dpair+antonymy%2E+In+Proceedings+of+the+2008+Conference+on+Empirical+Methods+in+Natural+Language+Processing%2C+pages+982%E2%80%93991%2C+Honolulu%2C+HI)
14. Raymond J. Mooney. 2007. Learning for semantic parsing. In Proceedings of the Eighth International Conference on Computational Linguistics and Intelligent Text Processing, pages 311–324, Mexico City, Mexico. Springer. [find](https://scholar.google.com/scholar?q=Raymond+J%2E+Mooney%2E+2007%2E+Learning+for+semantic+parsing%2E+In+Proceedings+of+the+Eighth+International+Conference+on+Computational+Linguistics+and+Intelligent+Text+Processing%2C+pages+311%E2%80%93324%2C+Mexico+City%2C+)
15. Hoifung Poon and Pedro Domingos. 2006. Sound and efficient inference with probabilistic and deterministic dependencies. In Proceedings of the Twenty First National Conference on Artificial Intelligence, pages 458–463, Boston, MA. AAAI Press. [find](https://scholar.google.com/scholar?q=Hoifung+Poon+and+Pedro+Domingos%2E+2006%2E+Sound+and+efficient+inference+with+probabilistic+and+deterministic+dependencies%2E+In+Proceedings+of+the+Twenty+First+National+Conference+on+Artificial+Intelligenc)
16. Hoifung Poon and Pedro Domingos. 2008. Joint unsupervised coreference resolution with Markov logic. In Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing, pages 649–658, Honolulu, HI. ACL. [find](https://scholar.google.com/scholar?q=Hoifung+Poon+and+Pedro+Domingos%2E+2008%2E+Joint+unsupervised+coreference+resolution+with+Markov+logic%2E+In+Proceedings+of+the+2008+Conference+on+Empirical+Methods+in+Natural+Language+Processing%2C+pages+649) [our card](https://md.research.elamur.ai/free/en/papers/poon-2008-joint-unsupervised-coreference-resolution-markov/)
17. M. Richardson and P. Domingos. 2006. Markov logic networks. Machine Learning, 62:107–136. [find](https://scholar.google.com/scholar?q=M%2E+Richardson+and+P%2E+Domingos%2E+2006%2E+Markov+logic+networks%2E+Machine+Learning%2C+62%3A107%E2%80%93136%2E)
18. Alexander Yates and Oren Etzioni. 2009. Unsupervised methods for determining object and relation synonyms on the web. Journal of Artificial Intelligence Research, 34:255–296. [find](https://scholar.google.com/scholar?q=Alexander+Yates+and+Oren+Etzioni%2E+2009%2E+Unsupervised+methods+for+determining+object+and+relation+synonyms+on+the+web%2E+Journal+of+Artificial+Intelligence+Research%2C+34%3A255%E2%80%93296%2E)
19. Luke S. Zettlemoyer and Michael Collins. 2005. Learning to map sentences to logical form: Structured classification with probabilistic categorial grammers. In Proceedings of the Twenty First Conference on Uncertainty in Artificial Intelligence, pages 658–666, Edinburgh, Scotland. AUAI Press. [find](https://scholar.google.com/scholar?q=Luke+S%2E+Zettlemoyer+and+Michael+Collins%2E+2005%2E+Learning+to+map+sentences+to+logical+form%3A+Structured+classification+with+probabilistic+categorial+grammers%2E+In+Proceedings+of+the+Twenty+First+Conferenc)
20. Luke S. Zettlemoyer and Michael Collins. 2007. Online learning of relaxed CCG grammars for parsing to logical form. In Proceedings of the Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, pages 878–887, Prague, Czech. ACL. [find](https://scholar.google.com/scholar?q=Luke+S%2E+Zettlemoyer+and+Michael+Collins%2E+2007%2E+Online+learning+of+relaxed+CCG+grammars+for+parsing+to+logical+form%2E+In+Proceedings+of+the+Joint+Conference+on+Empirical+Methods+in+Natural+Language+Proc)