---
title: "Information Fusion in the Context of — Barzilay 1999"
description: "The paper addresses multi-document summarization of news: many agencies report the same event, and picking whole sentences either repeats content or…"
authors: ["Regina Barzilay", "Kathleen R. McKeown", "Michael Elhadad"]
year: 1999
venue: "Proceedings of the 37th Annual Meeting of the Association for Computational Linguistics"
source: "https://aclanthology.org/P99-1071/"
doi: "10.3115/1034678.1034760"
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/barzilay-1999-information-fusion-context-multi-document/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/barzilay-1999-information-fusion-context-multi-document/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [1999](https://md.research.elamur.ai/free/en/papers/1999/) / Information Fusion in the Context of Multi-Document Summarization

# Information Fusion in the Context of Multi-Document Summarization

**Authors:** [Regina Barzilay](https://md.research.elamur.ai/free/en/authors/regina-barzilay/), [Kathleen R. McKeown](https://md.research.elamur.ai/free/en/authors/kathleen-r-mckeown/), [Michael Elhadad](https://md.research.elamur.ai/free/en/authors/michael-elhadad/) · **Year:** 1999 · **Venue:** Proceedings of the 37th Annual Meeting of the 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/P99-1071/) · [PDF](https://aclanthology.org/P99-1071.pdf) · [DOI 10.3115/1034678.1034760](https://doi.org/10.3115/1034678.1034760) · [OpenAlex](https://openalex.org/W2118733980)

**Citations:** 385 citations (openalex-snapshot · 2026-10-14) · [92](#cited-by) in our library, of the 140884 papers whose bibliographies we read

## Summary
The paper addresses multi-document summarization of news: many agencies report the same event, and picking whole sentences either repeats content or favors certain sources. The authors take as input "themes" (groups of similar sentences found by an earlier stage of a larger Columbia system) and build a new summary sentence from the material those sentences share. Each sentence is parsed with Collins' statistical parser and converted by rules into a DSYNT dependency tree. Pairs of trees are then compared recursively, using paraphrasing rules derived by hand from a training portion of a theme corpus. Phrases that recur often enough form the summary content. These are ordered, enriched with clarifying details, and realized as fluent text through FUF/SURGE. The fully implemented system produced, for example, a coherent summary from 12 Bosnia-related articles. The work matters because it was an early demonstration that generation, rather than extraction, can merge overlapping reports while cutting repetition and keeping sources balanced.

## Key points
- **Pipeline:**
  - Themes come from an earlier processing stage, described in work by Eskin et al. (1999) and by McKeown et al. (1999).
  - Sentences are parsed with the Collins (1996) parser.
  - A rule-based converter maps phrase-structure trees to DSYNT (Kittredge and Mel'čuk), dropping determiners and auxiliaries and storing lemmas plus grammatical features in each node.
  - Generation uses FUF/SURGE.
- **Intersection algorithm:**
  - The algorithm starts from verb-rooted trees and descends while nodes match.
  - A "full phrase" is a verb with at least two constituents.
  - Every pair of sentences in a theme is compared, and intersections whose frequency passes a threshold are kept.
- **Paraphrasing rules:**
  - The rules were collected manually from the training part of the theme corpus.
  - They target surface-level paraphrases, such as constituent reordering and omitting empty heads ("group of students" vs. "students").
  - Paraphrases that require world knowledge were set aside.
- **Content enrichment:** the content planner adds entity descriptions, temporal references and newswire source attributions. The sentence generator relies on surface features instead of semantic or pragmatic input.
- **Known failure:** the system keeps no surrounding context for an intersected phrase. When that context negated the matched clause, the summary became false; this happened once in the test cases.

## Interesting ideas
- **Treat repetition across sources as a salience signal.** Recurrence becomes a voting mechanism for content selection, which also keeps any single outlet from dominating. This suits aggregating Ukrainian news or trade reports from many sources.
- **Compare sentences as dependency trees with function words removed.** This makes matching robust to word order and inflection, which is promising for morphologically rich, free-word-order languages like Ukrainian, for example over UD trees.
- **Gather paraphrase rules from a training split before writing the matcher.** Studying which surface variations actually occur in the corpus focuses effort on the most frequent ones. The resulting rules can later serve as seeds for learning more.
- **Separate "what is shared" from "how to say it."** Selection works on predicate-argument structure and realization happens afterwards, so the output can be more concise than any single source sentence.
- **Check the context of extracted fragments for scope.** The negation failure shows that a phrase lifted out of a denying context can flip meaning. Any fusion or extraction system should verify such scope before outputting a fragment.

*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

92 papers of our library cite it (of the 140884 whose bibliographies we read); 6 of them have a page here, 86 not yet.

- [Sentence Fusion for Multidocument News Summarization](https://md.research.elamur.ai/free/en/papers/barzilay-2005-sentence-fusion-multidocument-news-summarization/)
- [Multi-News: A Large-Scale Multi-Document Summarization Dataset and Abstractive Hierarchical Model](https://md.research.elamur.ai/free/en/papers/fabbri-2019-multi-news-large-scale-multi/)
- [Annotating and Modeling Fine-grained Factuality in Summarization](https://md.research.elamur.ai/free/en/papers/goyal-2021-annotating-modeling-fine-grained-factuality/)
- [Abstract Meaning Representation for Multi-Document Summarization](https://md.research.elamur.ai/free/en/papers/liao-2018-abstract-meaning-representation-multi-document/)
- [Syntax-based Alignment of Multiple Translations: Extracting Paraphrases and Generating New Sentences](https://md.research.elamur.ai/free/en/papers/pang-2003-syntax-based-alignment-multiple-translations/)
- [Summarizing Scientific Articles: Experiments with Relevance and Rhetorical Status](https://md.research.elamur.ai/free/en/papers/teufel-2002-summarizing-scientific-articles-experiments-relevance/)

## Cite

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

```bibtex
@inproceedings{barzilay1999information,
  title = {{Information Fusion in the Context of Multi-Document Summarization}},
  author = {Barzilay, Regina and McKeown, Kathleen R. and Elhadad, Michael},
  booktitle = {Proceedings of the 37th Annual Meeting of the Association for Computational Linguistics},
  year = {1999},
  doi = {10.3115/1034678.1034760},
  url = {https://doi.org/10.3115/1034678.1034760}
}
```

- [← previous Improved Alignment Models for Statistical Machine Translation](https://md.research.elamur.ai/free/en/papers/och-1999-improved-alignment-models-statistical-machine/)
- [next → Interpreting and Generating Indirect Answers](https://md.research.elamur.ai/free/en/papers/green-1999-interpreting-generating-indirect-answers/)
- [↑ back to papers of 1999](https://md.research.elamur.ai/free/en/papers/1999/)

## References

18 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. James Allan, Jaime Carbonell, George Doddington, Jon Yamron, and Y. Yang. 1998. Topic detection and tracking pilot study: Final report. In Proceedings of the Broadcast News Understanding and Transcription Workshop, pages 194-218. [find](https://scholar.google.com/scholar?q=James+Allan%2C+Jaime+Carbonell%2C+George+Doddington%2C+Jon+Yamron%2C+and+Y%2E+Yang%2E+1998%2E+Topic+detection+and+tracking+pilot+study%3A+Final+report%2E+In+Proceedings+of+the+Broadcast+News+Understanding+and+Transcrip)
2. Regina Barzilay and Michael Elhadad. 1997. Using lexical chains for text summarization. In Proceedings of the A CL Workshop on Intelligent Scalable Text Summarization, pages 10-17, Madrid, Spain, August. ACL. [find](https://scholar.google.com/scholar?q=Regina+Barzilay+and+Michael+Elhadad%2E+1997%2E+Using+lexical+chains+for+text+summarization%2E+In+Proceedings+of+the+A+CL+Workshop+on+Intelligent+Scalable+Text+Summarization%2C+pages+10%2D17%2C+Madrid%2C+Spain%2C+Augu) [our card](https://md.research.elamur.ai/free/en/papers/barzilay-1997-lexical-chains-text-summarization/)
3. Jaime Carbonell and Jade Goldstein. 1998. The use of mmr, diversity-based reranking for reordering documents and producing summaries. In Proceedings of the 21st Annual International A CM SIGIR Conference on Research and Development in Information Retrieval, Melbourne, Australia, August. [find](https://scholar.google.com/scholar?q=Jaime+Carbonell+and+Jade+Goldstein%2E+1998%2E+The+use+of+mmr%2C+diversity%2Dbased+reranking+for+reordering+documents+and+producing+summaries%2E+In+Proceedings+of+the+21st+Annual+International+A+CM+SIGIR+Confere)
4. Michael Collins. 1996. A new statistical parser based on bigram lexical dependencies. In Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics, Santa Cruz, California. [find](https://scholar.google.com/scholar?q=Michael+Collins%2E+1996%2E+A+new+statistical+parser+based+on+bigram+lexical+dependencies%2E+In+Proceedings+of+the+35th+Annual+Meeting+of+the+Association+for+Computational+Linguistics%2C+Santa+Cruz%2C+California)
5. Mark Dras. 1997. Reluctant paraphrase: Textual restructuring under an optimisation model. In Proceedings of PA CLING97, pages 98-104, Ohme, Japan. [find](https://scholar.google.com/scholar?q=Mark+Dras%2E+1997%2E+Reluctant+paraphrase%3A+Textual+restructuring+under+an+optimisation+model%2E+In+Proceedings+of+PA+CLING97%2C+pages+98%2D104%2C+Ohme%2C+Japan%2E)
6. Michael Elhadad. 1993. Using Argumentation to Control Lexical Choice: A Functional Unification Implementation. Ph.D. thesis, Department of Computer Science, Columbia University, New York. [find](https://scholar.google.com/scholar?q=Michael+Elhadad%2E+1993%2E+Using+Argumentation+to+Control+Lexical+Choice%3A+A+Functional+Unification+Implementation%2E+Ph%2ED%2E+thesis%2C+Department+of+Computer+Science%2C+Columbia+University%2C+New+York%2E)
7. Eleazar Eskin, Judith Klavans, and Vasileios Hatzivassiloglou. 1999. Detecting similarity by apllying learning over indicators, submitted. [find](https://scholar.google.com/scholar?q=Eleazar+Eskin%2C+Judith+Klavans%2C+and+Vasileios+Hatzivassiloglou%2E+1999%2E+Detecting+similarity+by+apllying+learning+over+indicators%2C+submitted%2E)
8. Lidija Iordanskaja, Richard Kittredge, and Alain Polguere, 1991. Natural language Generation in Artificial Intelligence and Computational Linguistics, chapter 11. [find](https://scholar.google.com/scholar?q=Lidija+Iordanskaja%2C+Richard+Kittredge%2C+and+Alain+Polguere%2C+1991%2E+Natural+language+Generation+in+Artificial+Intelligence+and+Computational+Linguistics%2C+chapter+11%2E)
9. Kluwer Academic Publishers. Cristian Jacquemin, Judith L. Klavans, and Evelyne Tzoukermann. 1997. Expansion of multi-word terms for indexing and retrieval using morphology and syntax. In proceedings of the 35th Annual Meeting of the A CL, pages 24-31, Madrid, Spain, July. [find](https://scholar.google.com/scholar?q=Kluwer+Academic+Publishers%2E+Cristian+Jacquemin%2C+Judith+L%2E+Klavans%2C+and+Evelyne+Tzoukermann%2E+1997%2E+Expansion+of+multi%2Dword+terms+for+indexing+and+retrieval+using+morphology+and+syntax%2E+In+proceedings+o)
10. ACL. Hongyan Jing. 1999. Summary generation through intelligent cutting and pasting of the input document. PhD thesis proposal. [find](https://scholar.google.com/scholar?q=ACL%2E+Hongyan+Jing%2E+1999%2E+Summary+generation+through+intelligent+cutting+and+pasting+of+the+input+document%2E+PhD+thesis+proposal%2E)
11. Richard Kittredge and Igor A. Mel'Suk. 1983. Towards a computable model of meaning-text relations within a natural sublanguage. In Proceedings of the Eighth International Joint Conference on Artificial Intelligence (IJCAI-83), pages 657-659, Karlsruhe, West Germany, August. [find](https://scholar.google.com/scholar?q=Richard+Kittredge+and+Igor+A%2E+Mel%27Suk%2E+1983%2E+Towards+a+computable+model+of+meaning%2Dtext+relations+within+a+natural+sublanguage%2E+In+Proceedings+of+the+Eighth+International+Joint+Conference+on+Artificia)
12. Chin-Yew Lin and Eduard Hovy. 1997. Identifying topics by position. In Proceedings of the 5th A CL Conference on Applied Natural Language Processing, pages 283-290, Washington, D.C., April. [find](https://scholar.google.com/scholar?q=Chin%2DYew+Lin+and+Eduard+Hovy%2E+1997%2E+Identifying+topics+by+position%2E+In+Proceedings+of+the+5th+A+CL+Conference+on+Applied+Natural+Language+Processing%2C+pages+283%2D290%2C+Washington%2C+D%2EC%2E%2C+April%2E)
13. Inderjeet Mani and Eric Bloedorn. 1997. Multi-document summarization by graph search and matching. In Proceedings of the Fifteenth National Conference on Artificial Intelligence (AAAI-97), pages 622-628, Providence, Rhode Island. AAAI. [find](https://scholar.google.com/scholar?q=Inderjeet+Mani+and+Eric+Bloedorn%2E+1997%2E+Multi%2Ddocument+summarization+by+graph+search+and+matching%2E+In+Proceedings+of+the+Fifteenth+National+Conference+on+Artificial+Intelligence+%28AAAI%2D97%29%2C+pages+622%2D6)
14. Daniel Marcu. 1997. From discourse structures to text summaries. In Proceedings of the A CL Workshop on Intelligent Scalable Text Summarization, pages 82-88, Madrid, Spain, August. ACL. [find](https://scholar.google.com/scholar?q=Daniel+Marcu%2E+1997%2E+From+discourse+structures+to+text+summaries%2E+In+Proceedings+of+the+A+CL+Workshop+on+Intelligent+Scalable+Text+Summarization%2C+pages+82%2D88%2C+Madrid%2C+Spain%2C+August%2E+ACL%2E)
15. Kathleen R McKeown, Judith Klavans, Vasileios Hatzivassiloglou, Regina Barzilay, and Eleazar Eskin. 1999. Towards multi-document summarization by reformulation: Progress and prospects, submitted. [find](https://scholar.google.com/scholar?q=Kathleen+R+McKeown%2C+Judith+Klavans%2C+Vasileios+Hatzivassiloglou%2C+Regina+Barzilay%2C+and+Eleazar+Eskin%2E+1999%2E+Towards+multi%2Ddocument+summarization+by+reformulation%3A+Progress+and+prospects%2C+submitted%2E)
16. Dragomir R. Radev and Kathleen R. McKeown. 1998. Generating natural language summaries from multiple on-line sources. Computational Linguistics, 24(3):469-500, September. [find](https://scholar.google.com/scholar?q=Dragomir+R%2E+Radev+and+Kathleen+R%2E+McKeown%2E+1998%2E+Generating+natural+language+summaries+from+multiple+on%2Dline+sources%2E+Computational+Linguistics%2C+24%283%29%3A469%2D500%2C+September%2E)
17. Jacques Robin. 1994. Revision-Based Generation of Natural Language Summaries Pro-riding Historical Background: Corpus-Based Analysis, Design, Implementation, and Evaluation. Ph.D. thesis, Department of Computer Science, Columbia University, NY. [find](https://scholar.google.com/scholar?q=Jacques+Robin%2E+1994%2E+Revision%2DBased+Generation+of+Natural+Language+Summaries+Pro%2Driding+Historical+Background%3A+Corpus%2DBased+Analysis%2C+Design%2C+Implementation%2C+and+Evaluation%2E+Ph%2ED%2E+thesis%2C+Department+o)
18. Gerald Salton, James Allan, Chris Buckley, and Amit Singhal. 1991. Automatic analysis, theme generation, and summarization of machine-readable texts. Science, 264:1421-1426, June. [find](https://scholar.google.com/scholar?q=Gerald+Salton%2C+James+Allan%2C+Chris+Buckley%2C+and+Amit+Singhal%2E+1991%2E+Automatic+analysis%2C+theme+generation%2C+and+summarization+of+machine%2Dreadable+texts%2E+Science%2C+264%3A1421%2D1426%2C+June%2E)