---
title: "Learning the Fine-Grained Information Status of Discourse Entities"
description: "They build two systems on 147 annotated Switchboard dialogues: a hand-written… 78.7% accuracy. 20 references, cited by 19 papers here."
authors: ["Altaf Rahman", "Vincent Ng"]
year: 2012
venue: "Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics"
source: "https://aclanthology.org/E12-1081/"
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/rahman-2012-learning-fine-grained-information-status/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/rahman-2012-learning-fine-grained-information-status/)

[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/) / Learning the Fine-Grained Information Status of Discourse Entities

# Learning the Fine-Grained Information Status of Discourse Entities

**Authors:** [Altaf Rahman](https://md.research.elamur.ai/free/en/authors/altaf-rahman/) and [Vincent Ng](https://md.research.elamur.ai/free/en/authors/vincent-ng/) · **Year:** 2012 · **Venue:** [Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics](https://md.research.elamur.ai/free/en/venues/eacl/) ([EACL 2012](https://md.research.elamur.ai/free/en/venues/eacl/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/E12-1081/) · [PDF](https://aclanthology.org/E12-1081.pdf) · [OpenAlex](https://openalex.org/W1894881655)

**Citations:** 60 citations (OpenAlex, 2026-10-14) · [19](https://md.research.elamur.ai/free/en/papers/rahman-2012-learning-fine-grained-information-status/cited-by/) in our library, of the 141,840 papers whose bibliographies we read

## Summary
The paper addresses fine-grained information status (IS) determination. The task is to label each noun phrase in a dialogue with one of 16 subtypes from Nissim et al.'s (2004) scheme, which sorts entities into old, mediated (med) or new and then subdivides the first two. The authors argue that earlier work did poorly on new entities because it used only shallow string and syntactic cues. They therefore add semantic and world knowledge from knowledge bases, along with coreference chains.

They build two systems on 147 annotated Switchboard dialogues:
- a hand-written ruleset that combines these knowledge sources;
- a multi-class SVM whose features include the rules themselves and the rules' outputs.

With automatic (Stanford) coreference, the learned model reaches 78.7% accuracy, 21.3 points above the rules. The same features applied to the coarse 3-class task lift accuracy to 91.7%, and the F-score for new entities nearly doubles. This is reported as the first empirical study of automatic IS subtype prediction. The authors also suggest IS and coreference could be bootstrapped from each other.

## Key points
- Data: Nissim et al.'s (2004) Switchboard IS annotations, also released by the LDC in the NXT corpus. There are 117 training and 30 test dialogues, with 58,835 annotated NPs in total.
- The most frequent test classes are med/set (22.3%), new (16.2%), old/identity (15.8%) and old/general (14.2%). Classes such as med/func_value make up under 1%.
- Learned model accuracy is 78.7% with Stanford coreference and 86.4% with gold coreference. The rule-based baseline reaches 57.4% with Stanford coreference.
- Five feature groups are used: rule predictions, markable predictions, rule conditions, unigrams and markables. Removing any one group causes a significant drop (paired t-test, p \< 0.05). Unigrams are the strongest group used alone, and markables are the weakest.
- On the 3-class task, accuracy rises from 82.9% with R&N's features to 91.7%, and F-score on new rises from 46.7% to 87.2%. Rare, hard subtypes remain nearly unsolved: old/event scores 4.5 F and med/bound 5.1 F.

## Interesting ideas
- Hand-written rules are not used as a competing system; their firing conditions and their predicted labels are fed to the learner as features, and the learner discovers how to weight and combine them.
- A mention's discourse "newness" is predicted using coreference output, and the authors note that the two tasks could feed each other in a loop rather than forming a circular dependency.
- The mediated subtypes give a compact way to record how an unseen entity becomes inferable from earlier context: part-whole, set-subset, situation, an event-triggered reading, or the value of a function such as a temperature.
- Ties between valid labels are broken by a fixed precedence order over subtypes, so every NP gets exactly one label (e.g., a repeated proper name counts as old/identity, not med/general).
- Each knowledge source is evaluated twice: by removing it from the full system, and by training a classifier on that source alone. The two views reveal different things; for example, rule features work much better with gold than with automatic coreference.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (learn): Mova's self-induced repair rules could become features for a learned scorer instead of hard overrides, keeping each decision traceable to the rule that fired.
- **I2** (frames): the prag coreference step could pre-filter mentions labelled new so they are never resolved, cutting false links, while coreference chains in turn refine the new/old labels.
- **I3** (world): when a new entity enters the world of a text, it could store an explicit bridging relation to its anchor (e.g., door part-of house, 30 degrees value-of temperature), which makes the inference explainable.
- **I4** (world): a declared precedence over entity-status labels would give deterministic, inspectable decisions where several analyses are legitimate.
- **I5** (learn): ablating each Mova witness or knowledge source both ways, under gold and predicted upstream input, would show which components depend on perfect upstream analysis.

*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

- [The University of Texas at Dallas](https://md.research.elamur.ai/free/en/universities/the-university-of-texas-at-dallas/) — University, Richardson, United States · 47 works · 5,363 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 (1): [🧵 Discourse analysis](https://md.research.elamur.ai/free/en/topics/discourse-analysis/) · 394 works

## Cite

Cite as: Altaf Rahman and Vincent Ng. 2012. Learning the Fine-Grained Information Status of Discourse Entities. Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics.

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

```bibtex
@inproceedings{rahman2012learning,
  title = {{Learning the Fine-Grained Information Status of Discourse Entities}},
  author = {Altaf Rahman and Vincent Ng},
  booktitle = {Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics},
  year = {2012},
  url = {https://aclanthology.org/E12-1081/}
}
```

- [← previous Sentence Boundary Detection: A Long Solved Problem?](https://md.research.elamur.ai/free/en/papers/read-2012-sentence-boundary-detection-long-solved/)
- [next → Modality and Negation in SIMT Use of Modality and Negation in Semantically-Informed Syntactic MT](https://md.research.elamur.ai/free/en/papers/baker-2012-modality-negation-simt-use-modality/)
- [↑ back to papers of 2012](https://md.research.elamur.ai/free/en/papers/2012/)

## References

20 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Collin F. Baker, Charles J. Fillmore, and John B. Lowe. 1998. [The Berkeley FrameNet project](https://md.research.elamur.ai/free/en/papers/baker-1998-berkeley-framenet/). In Proceedings of the 36th Annual Meeting of the Association for Computational Linguistics and the 17th International Conference on Computational Linguistics, Volume 1, pages 86–90.
2. Sasha Calhoun, Jean Carletta, Jason Brenier, Neil Mayo, Dan Jurafsky, Mark Steedman, and David Beaver. 2010. The NXT-format Switchboard corpus: A rich resource for investigating the syntax, semantics, pragmatics and prosody of dialogue. Language Resources and Evaluation, 44(4):387–419. [↗](https://doi.org/10.1007/s10579-010-9120-1)
3. Miriam Eckert and Michael Strube. 2001. Dialogue acts, synchronising units and anaphora resolution. Journal of Semantics, 17(1):51–89. [find](https://scholar.google.com/scholar?q=Miriam+Eckert+and+Michael+Strube%2E+2001%2E+Dialogue+acts%2C+synchronising+units+and+anaphora+resolution%2E+Journal+of+Semantics%2C+17%281%29%3A51%E2%80%9389%2E)
4. Anthony Fader, Stephen Soderland, and Oren Etzioni. 2011. Identifying relations for open information extraction. In Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pages 1535–1545. [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+2011+Conference+on+Empirical+Methods+in+Natural+Language+Processi)
5. Christiane Fellbaum. 1998. WordNet: An Electronic Lexical Database. MIT Press, Cambridge, MA. [find](https://scholar.google.com/scholar?q=Christiane+Fellbaum%2E+1998%2E+WordNet%3A+An+Electronic+Lexical+Database%2E+MIT+Press%2C+Cambridge%2C+MA%2E)
6. Caroline Gasperin and Ted Briscoe. 2008. [Statistical anaphora resolution in biomedical texts](https://md.research.elamur.ai/free/en/papers/gasperin-2008-statistical-anaphora-resolution-biomedical-texts/). In Proceedings of the 22nd International Conference on Computational Linguistics, pages 257–264.
7. Michael Götze, Thomas Weskott, Cornelia Endriss, Ines Fiedler, Stefan Hinterwimmer, Svetlana Petrova, Anne Schwarz, Stavros Skopeteas, and Ruben Stoel. 2007. Information structure. In Working Papers of the SFB632, Interdisciplinary Studies on Information Structure (ISIS). Potsdam: Universitätsverlag Potsdam. [find](https://scholar.google.com/scholar?q=Michael+Go%CC%88tze%2C+Thomas+Weskott%2C+Cornelia+Endriss%2C+Ines+Fiedler%2C+Stefan+Hinterwimmer%2C+Svetlana+Petrova%2C+Anne+Schwarz%2C+Stavros+Skopeteas%2C+and+Ruben+Stoel%2E+2007%2E+Information+structure%2E+In+Working+Papers+)
8. Eva Hajičová. 1984. Topic and focus. In Contributions to Functional Syntax, Semantics, and Language Comprehension (LLSEE 16), pages 189–202. John Benjamins, Amsterdam. [↗](https://doi.org/10.1075/llsee.16.12haj)
9. Michael A. K. Halliday. 1976. Notes on transitivity and theme in English. Journal of Linguistics, 3(2):199–244. [find](https://scholar.google.com/scholar?q=Michael+A%2E+K%2E+Halliday%2E+1976%2E+Notes+on+transitivity+and+theme+in+English%2E+Journal+of+Linguistics%2C+3%282%29%3A199%E2%80%93244%2E)
10. Heeyoung Lee, Yves Peirsman, Angel Chang, Nathanael Chambers, Mihai Surdeanu, and Dan Jurafsky. 2011. [Stanford’s multi-pass sieve coreference resolution system at the CoNLL-2011 shared task](https://md.research.elamur.ai/free/en/papers/lee-2011-stanford-s-multi-pass-sieve/). In Proceedings of the Fifteenth Conference on Computational Natural Language Learning: Shared Task, pages 28–34.
11. Malvina Nissim, Shipra Dingare, Jean Carletta, and Mark Steedman. 2004. An annotation scheme for information status in dialogue. In Proceedings of the 4th International Conference on Language Resources and Evaluation, pages 1023–1026. [find](https://scholar.google.com/scholar?q=Malvina+Nissim%2C+Shipra+Dingare%2C+Jean+Carletta%2C+and+Mark+Steedman%2E+2004%2E+An+annotation+scheme+for+information+status+in+dialogue%2E+In+Proceedings+of+the+4th+International+Conference+on+Language+Resource)
12. Malvina Nissim. 2003. Annotation scheme for information status in dialogue. Available from [http://www.stanford.edu/class/cs224u/guidelines-infostatus.pdf](http://www.stanford.edu/class/cs224u/guidelines-infostatus.pdf). [↗](http://www.stanford.edu/class/cs224u/guidelines-infostatus.pdf)
13. Malvina Nissim. 2006. Learning information status of discourse entities. In Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing, pages 94–102. [find](https://scholar.google.com/scholar?q=Malvina+Nissim%2E+2006%2E+Learning+information+status+of+discourse+entities%2E+In+Proceedings+of+the+2006+Conference+on+Empirical+Methods+in+Natural+Language+Processing%2C+pages+94%E2%80%93102%2E)
14. Ellen F. Prince. 1981. Toward a taxonomy of given-new information. In P. Cole, editor, Radical Pragmatics, pages 223–255. New York, N.Y.: Academic Press. [find](https://scholar.google.com/scholar?q=Ellen+F%2E+Prince%2E+1981%2E+Toward+a+taxonomy+of+given%2Dnew+information%2E+In+P%2E+Cole%2C+editor%2C+Radical+Pragmatics%2C+pages+223%E2%80%93255%2E+New+York%2C+N%2EY%2E%3A+Academic+Press%2E)
15. Ellen F. Prince. 1992. The ZPG letter: Subjects, definiteness, and information-status. In Discourse Description: Diverse Analysis of a Fund Raising Text, pages 295–325. John Benjamins, Philadelphia/Amsterdam. [find](https://scholar.google.com/scholar?q=Ellen+F%2E+Prince%2E+1992%2E+The+ZPG+letter%3A+Subjects%2C+definiteness%2C+and+information%2Dstatus%2E+In+Discourse+Description%3A+Diverse+Analysis+of+a+Fund+Raising+Text%2C+pages+295%E2%80%93325%2E+John+Benjamins%2C+Philadelphia%2FAm)
16. Altaf Rahman and Vincent Ng. 2011. [Learning the information status of noun phrases in spoken dialogues](https://md.research.elamur.ai/free/en/papers/rahman-2011-learning-information-status-noun-phrases/). In Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pages 1069–1080.
17. Arndt Riester, David Lorenz, and Nina Seemann. 2010. A recursive annotation scheme for referential information status. In Proceedings of the Seventh International Conference on Language Resources and Evaluation, pages 717–722. [find](https://scholar.google.com/scholar?q=Arndt+Riester%2C+David+Lorenz%2C+and+Nina+Seemann%2E+2010%2E+A+recursive+annotation+scheme+for+referential+information+status%2E+In+Proceedings+of+the+Seventh+International+Conference+on+Language+Resources+and+)
18. Mark Steedman. 2000. The Syntactic Process. The MIT Press, Cambridge, MA. [↗](https://doi.org/10.7551/mitpress/6591.001.0001)
19. Ioannis Tsochantaridis, Thomas Hofmann, Thorsten Joachims, and Yasemin Altun. 2004. Support vector machine learning for interdependent and structured output spaces. In Proceedings of the 21st International Conference on Machine Learning, pages 104–112. [find](https://scholar.google.com/scholar?q=Ioannis+Tsochantaridis%2C+Thomas+Hofmann%2C+Thorsten+Joachims%2C+and+Yasemin+Altun%2E+2004%2E+Support+vector+machine+learning+for+interdependent+and+structured+output+spaces%2E+In+Proceedings+of+the+21st+Internat)
20. Enric Vallduvı́. 1992. The Informational Component. Garland, New York. [find](https://scholar.google.com/scholar?q=Enric+Vallduv%C4%B1%CC%81%2E+1992%2E+The+Informational+Component%2E+Garland%2C+New+York%2E)

*Updated 24 October 2026*