---
title: "Weakly Supervised Training of Semantic Parsers (2012)"
description: "Krishnamurthy and Mitchell show how to train a broad-coverage parser with no… 80% precision. 25 references, cited by 70 papers here."
authors: ["Jayant Krishnamurthy", "Tom Mitchell"]
year: 2012
venue: "Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning"
source: "https://aclanthology.org/D12-1069/"
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/krishnamurthy-2012-weakly-supervised-training-semantic-parsers/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/krishnamurthy-2012-weakly-supervised-training-semantic-parsers/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2012](https://md.research.elamur.ai/free/en/papers/2012/) / Weakly Supervised Training of Semantic Parsers

# Weakly Supervised Training of Semantic Parsers

**Authors:** [Jayant Krishnamurthy](https://md.research.elamur.ai/free/en/authors/jayant-krishnamurthy/) and [Tom Mitchell](https://md.research.elamur.ai/free/en/authors/tom-m-mitchell/) · **Year:** 2012 · **Venue:** [Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning](https://md.research.elamur.ai/free/en/venues/emnlp/) ([EMNLP 2012](https://md.research.elamur.ai/free/en/venues/emnlp/2012/)) · **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/D12-1069/) · [PDF](https://aclanthology.org/D12-1069.pdf)

**Citations:** [70](https://md.research.elamur.ai/free/en/papers/krishnamurthy-2012-weakly-supervised-training-semantic-parsers/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Semantic parsers usually need sentences hand-annotated with logical forms, which keeps them confined to narrow domains. Krishnamurthy and Mitchell show how to train a broad-coverage parser with no sentence-level labels. They use only a knowledge base (Freebase) and a web corpus processed by an off-the-shelf dependency parser. A probabilistic Combinatory Categorial Grammar (CCG), whose logical forms use 77 Freebase predicates, is trained against two weak constraints:
- **Semantic constraint:** every known fact about an entity pair must be expressed by at least one sentence mentioning both entities, and no parse may assert a fact absent from the knowledge base.
- **Syntactic constraint:** a parse's dependencies must be a subset of those in the automatic dependency parse.

Both constraints are encoded as deterministic factors in a factor graph built for each entity pair. Parameters are learned with beam-search-based approximate inference. The resulting parser matches a state-of-the-art weakly supervised relation extractor on binary relations. It also produces richer conjunctive logical forms, reaching 80% precision and 56% recall on natural-language queries against Freebase. This suggests that cheap, indirect supervision can stand in for annotated logical forms.

## Key points
- **Inputs:** a knowledge base K = (E, R, C, Δ) of entities, categories, relations and facts; dependency-parsed sentences; a CCG lexicon; and an entity-mention detector (for example, simple string matching).
- **Representation:** a log-linear CCG whose logical forms use 77 Freebase relations and categories, including conjunctions of predicates with shared arguments.
- **Semantic constraint:** this reuses the multiple deterministic-OR constraint of Hoffmann et al. (2011). The model is instantiated once per (e1, e2) tuple, which keeps the graphical model small.
- **Results:** binary relation extraction is on par with Hoffmann et al. (2011). On Freebase queries the parser reaches 80% precision and 56% recall. The PARSE vs. PARSE-DEP comparison confirms that syntactic supervision helps.
- **Error analysis and limitations:**
  - Main errors come from missing lexical entries for rare words (e.g., "ex-guitarist") and from confusing relations with similar type signatures, such as CITYLOCATEDINCOUNTRY vs. CITYCAPITALOFCOUNTRY.
  - The lexicon is initialized from hand-written dependency patterns, which the authors acknowledge as a limitation.

## Interesting ideas
- **Stacking weak signals:** neither the knowledge base nor the dependency parses alone would make learning well-posed, but together they constrain the parameter space enough. Combining complementary indirect signals is a general recipe for tasks that lack labeled data, such as low-resource language parsing.
- **Dependency parse as a soft upper bound:** the model treats the automatic parse as a ceiling on structure (the target parse must use a subset of its dependencies) rather than as a gold target. This lets existing UD-style parsers guide richer structured models without demanding exact agreement.
- **Group-level supervision via deterministic OR:** asking only that some sentence in the set of sentences mentioning an entity pair expresses each fact tolerates noisy alignment between texts and facts. It suits any distant-supervision setup built on a knowledge base or gazetteer, such as trade or geography databases.
- **Locality for tractability:** instantiating the model separately for each entity pair, instead of coupling all sentences, keeps inference manageable while preserving the constraint's spirit. This is a useful design pattern when a global constraint would explode model size.
- **Evaluation with executable queries:** testing the parser on natural-language queries against the knowledge base checks whether the full logical structure is recovered, not just binary relations. This makes it a stronger benchmark than relation extraction scores alone.

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

📚 Shelves (2): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works

🧭 Topics (1): [🧾 Semantic parsing & meaning representation](https://md.research.elamur.ai/free/en/topics/semantic-parsing/) · 185 works

## Cite

Cite as: Jayant Krishnamurthy and Tom Mitchell. 2012. Weakly Supervised Training of Semantic Parsers. Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning.

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

```bibtex
@inproceedings{krishnamurthy2012weakly,
  title = {{Weakly Supervised Training of Semantic Parsers}},
  author = {Jayant Krishnamurthy and Tom Mitchell},
  booktitle = {Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning},
  year = {2012},
  url = {https://aclanthology.org/D12-1069/}
}
```

- [← previous Learning-based Multi-Sieve Co-reference Resolution with Knowledge](https://md.research.elamur.ai/free/en/papers/ratinov-2012-learning-based-multi-sieve-co/)
- [next → Identifying hypernyms in distributional semantic spaces](https://md.research.elamur.ai/free/en/papers/lenci-2012-identifying-hypernyms-distributional-semantic-spaces/)
- [↑ back to papers of 2012](https://md.research.elamur.ai/free/en/papers/2012/)

## References

25 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Kurt Bollacker, Colin Evans, Praveen Paritosh, Tim Sturge, and Jamie Taylor. 2008. Freebase: a collaboratively created graph database for structuring human knowledge. In Proceedings of the 2008 ACM SIGMOD International Conference on Management of Data, pages 1247–1250. [find](https://scholar.google.com/scholar?q=Kurt+Bollacker%2C+Colin+Evans%2C+Praveen+Paritosh%2C+Tim+Sturge%2C+and+Jamie+Taylor%2E+2008%2E+Freebase%3A+a+collaboratively+created+graph+database+for+structuring+human+knowledge%2E+In+Proceedings+of+the+2008+ACM+SI)
2. Andrew Carlson, Justin Betteridge, Bryan Kisiel, Burr Settles, Estevam R. Hruschka Jr., and Tom M. Mitchell. 2010. Toward an architecture for never-ending language learning. In Proceedings of the Twenty-Fourth AAAI Conference on Artificial Intelligence. [find](https://scholar.google.com/scholar?q=Andrew+Carlson%2C+Justin+Betteridge%2C+Bryan+Kisiel%2C+Burr+Settles%2C+Estevam+R%2E+Hruschka+Jr%2E%2C+and+Tom+M%2E+Mitchell%2E+2010%2E+Toward+an+architecture+for+never%2Dending+language+learning%2E+In+Proceedings+of+the+Twen)
3. Stephen Clark and James R. Curran. 2007. [Wide-coverage efficient statistical parsing with CCG and log-linear models](https://md.research.elamur.ai/free/en/papers/clark-2007-wide-coverage-efficient-statistical-parsing/). Computational Linguistics, 33(4):493–552.
4. James Clarke, Dan Goldwasser, Ming-Wei Chang, and Dan Roth. 2010. Driving semantic parsing from the world’s response. In Proceedings of the Fourteenth Conference on Computational Natural Language Learning. [find](https://scholar.google.com/scholar?q=James+Clarke%2C+Dan+Goldwasser%2C+Ming%2DWei+Chang%2C+and+Dan+Roth%2E+2010%2E+Driving+semantic+parsing+from+the+world%E2%80%99s+response%2E+In+Proceedings+of+the+Fourteenth+Conference+on+Computational+Natural+Language+Lear)
5. 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 Conference on Empirical Methods in Natural Language Processing.
6. Mark Craven and Johan Kumlien. 1999. Constructing biological knowledge bases by extracting information from text sources. In Proceedings of the Seventh International Conference on Intelligent Systems for Molecular Biology. [find](https://scholar.google.com/scholar?q=Mark+Craven+and+Johan+Kumlien%2E+1999%2E+Constructing+biological+knowledge+bases+by+extracting+information+from+text+sources%2E+In+Proceedings+of+the+Seventh+International+Conference+on+Intelligent+Systems+)
7. Ruifang Ge and Raymond J. Mooney. 2005. A statistical semantic parser that integrates syntax and semantics. In Proceedings of the Ninth Conference on Computational Natural Language Learning. [find](https://scholar.google.com/scholar?q=Ruifang+Ge+and+Raymond+J%2E+Mooney%2E+2005%2E+A+statistical+semantic+parser+that+integrates+syntax+and+semantics%2E+In+Proceedings+of+the+Ninth+Conference+on+Computational+Natural+Language+Learning%2E)
8. Dan Goldwasser, Roi Reichart, James Clarke, and Dan Roth. 2011. Confidence driven unsupervised semantic parsing. Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Dan+Goldwasser%2C+Roi+Reichart%2C+James+Clarke%2C+and+Dan+Roth%2E+2011%2E+Confidence+driven+unsupervised+semantic+parsing%2E+Proceedings+of+the+49th+Annual+Meeting+of+the+Association+for+Computational+Linguistics)
9. Raphael Hoffmann, Congle Zhang, Xiao Ling, Luke S. Zettlemoyer, and Daniel S. Weld. 2011. Knowledge-based weak supervision for information extraction of overlapping relations. In The 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies. [find](https://scholar.google.com/scholar?q=Raphael+Hoffmann%2C+Congle+Zhang%2C+Xiao+Ling%2C+Luke+S%2E+Zettlemoyer%2C+and+Daniel+S%2E+Weld%2E+2011%2E+Knowledge%2Dbased+weak+supervision+for+information+extraction+of+overlapping+relations%2E+In+The+49th+Annual+Meeti)
10. Rohit J. Kate, Yuk Wah Wong, and Raymond J. Mooney. 2005. Learning to transform natural to formal languages. In Proceedings, The Twentieth National Conference on Artificial Intelligence and the Seventeenth Innovative Applications of Artificial Intelligence Conference. [find](https://scholar.google.com/scholar?q=Rohit+J%2E+Kate%2C+Yuk+Wah+Wong%2C+and+Raymond+J%2E+Mooney%2E+2005%2E+Learning+to+transform+natural+to+formal+languages%2E+In+Proceedings%2C+The+Twentieth+National+Conference+on+Artificial+Intelligence+and+the+Sevent)
11. Tom Kwiatkowski, Luke Zettlemoyer, Sharon Goldwater, and Mark Steedman. 2010. Inducing probabilistic CCG grammars from logical form with higher-order unification. In Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing. [find](https://scholar.google.com/scholar?q=Tom+Kwiatkowski%2C+Luke+Zettlemoyer%2C+Sharon+Goldwater%2C+and+Mark+Steedman%2E+2010%2E+Inducing+probabilistic+CCG+grammars+from+logical+form+with+higher%2Dorder+unification%2E+In+Proceedings+of+the+2010+Conference)
12. Percy Liang, Michael I. Jordan, and Dan Klein. 2011. Learning dependency-based compositional semantics. In Proceedings of the Association for Computational Linguistics, Portland, Oregon. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Percy+Liang%2C+Michael+I%2E+Jordan%2C+and+Dan+Klein%2E+2011%2E+Learning+dependency%2Dbased+compositional+semantics%2E+In+Proceedings+of+the+Association+for+Computational+Linguistics%2C+Portland%2C+Oregon%2E+Association+f)
13. Wei Lu, Hwee Tou Ng, Wee Sun Lee, and Luke S. Zettlemoyer. 2008. [A generative model for parsing natural language to meaning representations](https://md.research.elamur.ai/free/en/papers/lu-2008-generative-model-parsing-natural-language/). In Proceedings of the Conference on Empirical Methods in Natural Language Processing.
14. Mike Mintz, Steven Bills, Rion Snow, and Dan Jurafsky. 2009. [Distant supervision for relation extraction without labeled data](https://md.research.elamur.ai/free/en/papers/mintz-2009-distant-supervision-relation-extraction-without/). In Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP.
15. Joakim Nivre, Johan Hall, and Jens Nilsson. 2006. [Maltparser: A data-driven parser-generator for dependency parsing](https://md.research.elamur.ai/free/en/papers/nivre-2006-maltparser-data-driven-parser-generator/). In Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics.
16. Hoifung Poon and Pedro Domingos. 2009. [Unsupervised semantic parsing](https://md.research.elamur.ai/free/en/papers/poon-2009-unsupervised-semantic-parsing/). In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing.
17. Sebastian Riedel, Limin Yao, and Andrew McCallum. 2010. Modeling relations and their mentions without labeled text. In Proceedings of the 2010 European conference on Machine learning and Knowledge Discovery in Databases. [find](https://scholar.google.com/scholar?q=Sebastian+Riedel%2C+Limin+Yao%2C+and+Andrew+McCallum%2E+2010%2E+Modeling+relations+and+their+mentions+without+labeled+text%2E+In+Proceedings+of+the+2010+European+conference+on+Machine+learning+and+Knowledge+Dis)
18. Mark Steedman. 1996. Surface Structure and Interpretation. The MIT Press. [find](https://scholar.google.com/scholar?q=Mark+Steedman%2E+1996%2E+Surface+Structure+and+Interpretation%2E+The+MIT+Press%2E)
19. Fabian M. Suchanek, Gjergji Kasneci, and Gerhard Weikum. 2007. Yago: a core of semantic knowledge. In Proceedings of the 16th international conference on World Wide Web, WWW ’07, pages 697–706, New York, NY, USA. ACM. [find](https://scholar.google.com/scholar?q=Fabian+M%2E+Suchanek%2C+Gjergji+Kasneci%2C+and+Gerhard+Weikum%2E+2007%2E+Yago%3A+a+core+of+semantic+knowledge%2E+In+Proceedings+of+the+16th+international+conference+on+World+Wide+Web%2C+WWW+%E2%80%9907%2C+pages+697%E2%80%93706%2C+New+Yo)
20. Yuk Wah Wong and Raymond J. Mooney. 2006. [Learning for semantic parsing with statistical machine translation](https://md.research.elamur.ai/free/en/papers/wong-2006-learning-semantic-parsing-statistical-machine/). In Proceedings of the Human Language Technology Conference of the NAACL.
21. Yuk Wah Wong and Raymond J. Mooney. 2007. [Learning synchronous grammars for semantic parsing with lambda calculus](https://md.research.elamur.ai/free/en/papers/wong-2007-learning-synchronous-grammars-semantic-parsing/). In Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics.
22. Fei Wu and Daniel S. Weld. 2010. [Open information extraction using Wikipedia](https://md.research.elamur.ai/free/en/papers/wu-2010-open-information-extraction-wikipedia/). In Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics.
23. John M. Zelle and Raymond J. Mooney. 1996. [Learning to parse database queries using inductive logic programming](https://md.research.elamur.ai/free/en/papers/zelle-1996-learning-parse-database-queries-using/). In Proceedings of the thirteenth national conference on Artificial Intelligence.
24. Luke S. Zettlemoyer and Michael Collins. 2005. [Learning to map sentences to logical form: structured classification with probabilistic categorial grammars](https://md.research.elamur.ai/free/en/papers/zettlemoyer-2012-learning-map-sentences-logical-form/). In UAI ’05, Proceedings of the 21st Conference in Uncertainty in Artificial Intelligence.
25. Luke S. Zettlemoyer and Michael Collins. 2007. Online learning of relaxed ccg grammars for parsing to logical form. In Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning. [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+2007+Joint+Conference+on+Empirical+Methods+in+Natural+Language)

*Updated 29 October 2026*