---
title: "Semantic Parsing for Single-Relation Question — Yih 2014"
description: "The paper tackles open-domain question answering for single-relation factual questions, such as asking who founded a company."
authors: ["Wen-tau Yih", "Xiaodong He", "Christopher Meek"]
year: 2014
venue: "Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)"
source: "https://aclanthology.org/P14-2105/"
doi: "10.3115/v1/p14-2105"
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/yih-2014-semantic-parsing-single-relation-question/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/yih-2014-semantic-parsing-single-relation-question/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2014](https://md.research.elamur.ai/free/en/papers/2014/page/2/) / Semantic Parsing for Single-Relation Question Answering

# Semantic Parsing for Single-Relation Question Answering

**Authors:** [Wen-tau Yih](https://md.research.elamur.ai/free/en/authors/wen-tau-yih/), [Xiaodong He](https://md.research.elamur.ai/free/en/authors/xiaodong-he/), [Christopher Meek](https://md.research.elamur.ai/free/en/authors/christopher-meek/) · **Year:** 2014 · **Venue:** Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) · **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/P14-2105/) · [PDF](https://aclanthology.org/P14-2105.pdf) · [DOI 10.3115/v1/p14-2105](https://doi.org/10.3115/v1/p14-2105) · [OpenAlex](https://openalex.org/W2251143283)

**Citations:** 410 citations (openalex-snapshot · 2026-10-14) · [73](#cited-by) in our library, of the 140884 papers whose bibliographies we read

## Summary
The paper tackles open-domain question answering for single-relation factual questions, such as asking who founded a company. These can be answered with one KB lookup once the relation and the entity are correctly identified. The main obstacle is paraphrase: the same request can be worded in many ways. The authors split each question into an entity mention (a contiguous word span) and a relation pattern (the question with that span replaced by a placeholder). They try every possible split. Two convolutional semantic similarity models (CNNSM) then score how well a pattern fits a KB relation and how well a mention fits a KB entity. Both models are trained on parallel data derived from the PARALEX resources: WikiAnswers paraphrases and ReVerb triples. Each candidate triple is scored by multiplying the two model probabilities, and the best triple supplies the answer. On the PARALEX test set, the full system beats the rule-based PARALEX at every recall level, with a best F1 of 0.61 against 0.54. This suggests that learned similarity can replace hand-induced lexical rules for this kind of semantic parsing.

## Key points
- **Training data.** The PARALEX data contains about 1.8 million question–query pairs. From it, the authors extracted roughly 1.2 million pattern–relation pairs by exact surface matching of entities. Applying those patterns back to the questions revealed about 160 thousand additional mentions whose surface forms differ from the entity names.
- **Test data.** The evaluation uses 698 questions from 37 clusters. Answer triples that PARALEX had never returned were labeled by the authors themselves.
- **Model.** CNNSM represents words with letter-trigram word hashing, applies a convolution over a word window, and then max-pools across the sequence over 500 feature dimensions. A tanh semantic layer follows, and cosine similarity is computed between the resulting vectors. Training uses a softmax over 1 positive and 100 random negatives, with γ = 5, optimized with SGD.
- **Inference.** For each candidate pattern, the system keeps the top 150 relations. It then scores the entity arguments of every KB triple with those relations against the mention.
- **Results (Table 1).** Table 1 reports each system at a single operating point:

  | System | Description | Precision | Recall | F1 | MAP |
  |---|---|---|---|---|---|
  | CNNSM_pm | full system | 0.58 | 0.57 | 0.57 | 0.28 |
  | CNNSM_p | exact-match entities only | 0.61 | 0.49 | 0.54 | 0.20 |
  | PARALEX | rule-based baseline | 0.77 | 0.42 | 0.54 | 0.22 |

  Across thresholds, the best attainable F1 values are 0.61, 0.56 and 0.54, respectively.

## Interesting ideas
- **Splitting a question into mention and pattern.** This turns semantic parsing into two separate similarity-matching problems. The design could transfer to any lookup-style QA over a triple store, including small domain KBs such as geography or trade data.
- **Letter-trigram word hashing.** It keeps the input vocabulary small and copes with unseen words and misspellings. This is attractive for morphologically rich languages like Ukrainian or for noisy user-generated text.
- **Bootstrapping training pairs from weak alignment.** Exact string matches yield the first set of patterns. Those patterns are then reapplied to the questions to find new, non-identical mentions. This cheaply grows a parallel corpus without manual annotation.
- **Softmax over sampled negatives.** Cosine scores are pushed through a softmax against 100 random alternatives. This gives probabilities that can be multiplied across independent components, a simple recipe for any retrieval or linking model.
- **Comparing on precision–recall curves.** Evaluating across score thresholds, rather than at one operating point, shows trade-offs that a single F1 number hides. The ablation without fuzzy entity matching illustrates this: it equals the full system at high precision but falls behind as recall grows.

*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/)

## Cited by

Cited by

73 papers of our library cite it (of the 140884 whose bibliographies we read); 4 of them have a page here, 69 not yet.

- [Convolutional Neural Networks for Sentence Classification](https://md.research.elamur.ai/free/en/papers/kim-2014-convolutional-neural-networks-sentence-classification/)
- [Event Detection and Domain Adaptation with Convolutional Neural Networks](https://md.research.elamur.ai/free/en/papers/nguyen-2015-event-detection-domain-adaptation-convolutional/)
- [Relation Extraction: Perspective from Convolutional Neural Networks](https://md.research.elamur.ai/free/en/papers/nguyen-2015-relation-extraction-perspective-convolutional-neural/)
- [Question Answering on Freebase via Relation Extraction and Textual Evidence](https://md.research.elamur.ai/free/en/papers/xu-2016-question-answering-freebase-relation-extraction/)

## Cite

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

```bibtex
@inproceedings{yih2014semantic,
  title = {{Semantic Parsing for Single-Relation Question Answering}},
  author = {Yih, Wen-tau and He, Xiaodong and Meek, Christopher},
  booktitle = {Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)},
  year = {2014},
  doi = {10.3115/v1/p14-2105},
  url = {https://doi.org/10.3115/v1/p14-2105}
}
```

- [← previous Scoring Coreference Partitions of Predicted Mentions: A Reference Implementation](https://md.research.elamur.ai/free/en/papers/pradhan-2014-scoring-coreference-partitions-predicted-mentions/)
- [next → SemEval-2014 Task 10: Multilingual Semantic Textual Similarity](https://md.research.elamur.ai/free/en/papers/agirre-2014-semeval-2014-task-10-multilingual/)
- [↑ back to papers of 2014](https://md.research.elamur.ai/free/en/papers/2014/page/2/)

## References

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

1. Jonathan Berant, Andrew Chou, Roy Frostig, and Percy Liang. 2013. Semantic parsing on Freebase from question-answer pairs. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1533–1544, Seattle, Washington, USA, October. Association for Computational Ling… [find](https://scholar.google.com/scholar?q=Jonathan+Berant%2C+Andrew+Chou%2C+Roy+Frostig%2C+and+Percy+Liang%2E+2013%2E+Semantic+parsing+on+Freebase+from+question%2Danswer+pairs%2E+In+Proceedings+of+the+2013+Conference+on+Empirical+Methods+in+Natural+Languag)
2. David M Blei, Andrew Y Ng, and Michael I Jordan. 2003. Latent dirichlet allocation. the Journal of machine Learning research, 3:993–1022. [find](https://scholar.google.com/scholar?q=David+M+Blei%2C+Andrew+Y+Ng%2C+and+Michael+I+Jordan%2E+2003%2E+Latent+dirichlet+allocation%2E+the+Journal+of+machine+Learning+research%2C+3%3A993%E2%80%931022%2E) [our card](https://md.research.elamur.ai/free/en/papers/blei-2003-latent-dirichlet-allocation/)
3. Qingqing Cai and Alexander Yates. 2013. Large-scale semantic parsing via schema matching and lexicon extension. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 423–433, Sofia, Bulgaria, August. Association for Computational Li… [find](https://scholar.google.com/scholar?q=Qingqing+Cai+and+Alexander+Yates%2E+2013%2E+Large%2Dscale+semantic+parsing+via+schema+matching+and+lexicon+extension%2E+In+Proceedings+of+the+51st+Annual+Meeting+of+the+Association+for+Computational+Linguisti)
4. Ronan Collobert, Jason Weston, Leon Bottou, Michael Karlen, Koray Kavukcuoglu, and Pavel Kuksa. 2011. Natural language processing (almost) from scratch. Journal of Machine Learning Research. [find](https://scholar.google.com/scholar?q=Ronan+Collobert%2C+Jason+Weston%2C+Leon+Bottou%2C+Michael+Karlen%2C+Koray+Kavukcuoglu%2C+and+Pavel+Kuksa%2E+2011%2E+Natural+language+processing+%28almost%29+from+scratch%2E+Journal+of+Machine+Learning+Research%2E) [our card](https://md.research.elamur.ai/free/en/papers/collobert-2011-natural-language-processing-almost-scratch/)
5. Scott Deerwester, Susan Dumais, Thomas Landauer, George Furnas, and Richard Harshman. 1990. Indexing by latent semantic analysis. Journal of the American Society for Information Science, 41(6). [find](https://scholar.google.com/scholar?q=Scott+Deerwester%2C+Susan+Dumais%2C+Thomas+Landauer%2C+George+Furnas%2C+and+Richard+Harshman%2E+1990%2E+Indexing+by+latent+semantic+analysis%2E+Journal+of+the+American+Society+for+Information+Science%2C+41%286%29%2E)
6. Anthony Fader, Stephen Soderland, and Oren Etzioni. 2011. Identifying relations for open information extraction. In Proceedings of the Conference of Empirical Methods in Natural Language Processing (EMNLP ’11), Edinburgh, Scotland, UK, July 27-31. [find](https://scholar.google.com/scholar?q=Anthony+Fader%2C+Stephen+Soderland%2C+and+Oren+Etzioni%2E+2011%2E+Identifying+relations+for+open+information+extraction%2E+In+Proceedings+of+the+Conference+of+Empirical+Methods+in+Natural+Language+Processing+%28E)
7. Anthony Fader, Luke Zettlemoyer, and Oren Etzioni. 2013. Paraphrase-driven learning for open question answering. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1608–1618, Sofia, Bulgaria, August. Association for Computational… [find](https://scholar.google.com/scholar?q=Anthony+Fader%2C+Luke+Zettlemoyer%2C+and+Oren+Etzioni%2E+2013%2E+Paraphrase%2Ddriven+learning+for+open+question+answering%2E+In+Proceedings+of+the+51st+Annual+Meeting+of+the+Association+for+Computational+Linguist) [our card](https://md.research.elamur.ai/free/en/papers/fader-2013-paraphrase-driven-learning-open-question/)
8. Po-Sen Huang, Xiaodong He, Jianfeng Gao, Li Deng, Alex Acero, and Larry Heck. 2013. Learning deep structured semantic models for web search using clickthrough data. In Proceedings of the 22nd ACM international conference on Conference on information & knowledge management, pages 2333–2338. ACM. [find](https://scholar.google.com/scholar?q=Po%2DSen+Huang%2C+Xiaodong+He%2C+Jianfeng+Gao%2C+Li+Deng%2C+Alex+Acero%2C+and+Larry+Heck%2E+2013%2E+Learning+deep+structured+semantic+models+for+web+search+using+clickthrough+data%2E+In+Proceedings+of+the+22nd+ACM+inte)
9. Tom Kwiatkowski, Eunsol Choi, Yoav Artzi, and Luke Zettlemoyer. 2013. Scaling semantic parsers with on-the-fly ontology matching. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1545–1556, Seattle, Washington, USA, October. Association for Computation… [find](https://scholar.google.com/scholar?q=Tom+Kwiatkowski%2C+Eunsol+Choi%2C+Yoav+Artzi%2C+and+Luke+Zettlemoyer%2E+2013%2E+Scaling+semantic+parsers+with+on%2Dthe%2Dfly+ontology+matching%2E+In+Proceedings+of+the+2013+Conference+on+Empirical+Methods+in+Natural+)
10. Percy Liang, Michael I Jordan, and Dan Klein. 2013. Learning dependency-based compositional semantics. Computational Linguistics, 39(2):389–446. [find](https://scholar.google.com/scholar?q=Percy+Liang%2C+Michael+I+Jordan%2C+and+Dan+Klein%2E+2013%2E+Learning+dependency%2Dbased+compositional+semantics%2E+Computational+Linguistics%2C+39%282%29%3A389%E2%80%93446%2E)
11. Grégoire Mesnil, Xiaodong He, Li Deng, and Yoshua Bengio. 2013. Investigation of recurrent-neural-network architectures and learning methods for spoken language understanding. In Interspeech. [find](https://scholar.google.com/scholar?q=Gre%CC%81goire+Mesnil%2C+Xiaodong+He%2C+Li+Deng%2C+and+Yoshua+Bengio%2E+2013%2E+Investigation+of+recurrent%2Dneural%2Dnetwork+architectures+and+learning+methods+for+spoken+language+understanding%2E+In+Interspeech%2E)
12. Ruslan Salakhutdinov and Geoffrey Hinton. 2009. Semantic hashing. International Journal of Approximate Reasoning, 50(7):969–978. [find](https://scholar.google.com/scholar?q=Ruslan+Salakhutdinov+and+Geoffrey+Hinton%2E+2009%2E+Semantic+hashing%2E+International+Journal+of+Approximate+Reasoning%2C+50%287%29%3A969%E2%80%93978%2E)
13. Gerard Salton and Michael J. McGill. 1983. Introduction to Modern Information Retrieval. McGraw Hill. [find](https://scholar.google.com/scholar?q=Gerard+Salton+and+Michael+J%2E+McGill%2E+1983%2E+Introduction+to+Modern+Information+Retrieval%2E+McGraw+Hill%2E)
14. Yelong Shen, Xiaodong He, Jianfeng Gao, Li Deng, and Grégoire Mesnil. 2014a. A convolutional latent semantic model for web search. Technical Report MSR-TR-2014-55, Microsoft Research. [find](https://scholar.google.com/scholar?q=Yelong+Shen%2C+Xiaodong+He%2C+Jianfeng+Gao%2C+Li+Deng%2C+and+Gre%CC%81goire+Mesnil%2E+2014a%2E+A+convolutional+latent+semantic+model+for+web+search%2E+Technical+Report+MSR%2DTR%2D2014%2D55%2C+Microsoft+Research%2E)
15. Yelong Shen, Xiaodong He, Jianfeng Gao, Li Deng, and Grégoire Mesnil. 2014b. Learning semantic representations using convolutional neural networks for web search. In Proceedings of the Companion Publication of the 23rd International Conference on World Wide Web Companion, pages 373–374. [find](https://scholar.google.com/scholar?q=Yelong+Shen%2C+Xiaodong+He%2C+Jianfeng+Gao%2C+Li+Deng%2C+and+Gre%CC%81goire+Mesnil%2E+2014b%2E+Learning+semantic+representations+using+convolutional+neural+networks+for+web+search%2E+In+Proceedings+of+the+Companion+Publ)
16. Lappoon Tang and Raymond Mooney. 2001. Using multiple clause constructors in inductive logic programming for semantic parsing. In Machine Learning: ECML 2001, pages 466–477. Springer. [find](https://scholar.google.com/scholar?q=Lappoon+Tang+and+Raymond+Mooney%2E+2001%2E+Using+multiple+clause+constructors+in+inductive+logic+programming+for+semantic+parsing%2E+In+Machine+Learning%3A+ECML+2001%2C+pages+466%E2%80%93477%2E+Springer%2E)
17. Gokhan Tur, Li Deng, Dilek Hakkani-Tur, and Xiaodong He. 2012. Towards deeper understanding: deep convex networks for semantic utterance classification. In Acoustics, Speech and Signal Processing (ICASSP), 2012 IEEE International Conference on, pages 5045–5048. IEEE. [find](https://scholar.google.com/scholar?q=Gokhan+Tur%2C+Li+Deng%2C+Dilek+Hakkani%2DTur%2C+and+Xiaodong+He%2E+2012%2E+Towards+deeper+understanding%3A+deep+convex+networks+for+semantic+utterance+classification%2E+In+Acoustics%2C+Speech+and+Signal+Processing+%28ICA)
18. John Zelle and Raymond Mooney. 1996. Learning to parse database queries using inductive logic programming. In Proceedings of the National Conference on Artificial Intelligence, pages 1050–1055. [find](https://scholar.google.com/scholar?q=John+Zelle+and+Raymond+Mooney%2E+1996%2E+Learning+to+parse+database+queries+using+inductive+logic+programming%2E+In+Proceedings+of+the+National+Conference+on+Artificial+Intelligence%2C+pages+1050%E2%80%931055%2E)