Daniel Marcu. From Discourse Structures to Text Summaries. In Proceedings of the Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, August 1997. Association for Computational Linguistics.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....al [10] suggested that the binary precision and recall measures are too sensitive to provide a robust evaluation, and, for the task based analysis, the measures used do not translate to any indicative measure of quality. Among the various attempts to improve these evaluation techniques, D. Marcu [14] emphasizes the need for evaluating not only the actual output (summaries) but the underlying assumptions of the method as well. In this paper, extending Marcu s position, we replace the a posteriori human judgment by an a priori judgment, and evaluate summaries by the amount and level of detail ....

Daniel Marcu. From discourse structures to text summaries. In I. Mani and M. Maybury, editors, Proceedings of the ACL/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82 88, Madrid, Spain, 1997. UNED.

....are domain independent. CST posits that by identifying these cross document links , one can produce superior multi document summaries. The concept of using CST for multi document summaries relates to the that of using Rhetorical Structure Theory (RST) 1] for singledocument summarization [2]. However, while Marcu relied on cue phrases in implementing algorithms to discover the valid RST trees for a single document, such a technique is not very plausible for discovering CST links between documents. For instance, the cue phrase although statementX, statementY might indicate ....

Daniel Marcu. From Discourse Structures to Text Summaries. In The Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, July 11 1997.

....Luhn s work at IBM in the fifties [12] Most of the work in sentence extraction applied statistical techniques (frequency analysis, variance analysis, etc. to linguistic units suchastokens, names, anaphora, etc. e.g. 27, 19, 9, 18, 2] Other approaches include the utility of discourse structure [14], the combination of information extraction and language generation [11, 17, 24, 21, 16] and using machine learning to find patterns in text [28, 4, 26] Several researchers have extended various aspects of the single document approaches to look at multi documentsummarization [13, 21, 3, 7, ....

Daniel Marcu. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, 1997.

....operational environment, even very simple heuristics such as, for instance, take the first sentence from each segment have remarkably noticeable impact. In essence, this paper argues that a lexical repetition based model of linear segmentation offers highly 2 As opposed to hierarchical; see (Marcu, 1997). plausible schemes for deriving sentence based summaries with certain discourse properties, as a result improving upon an already respectable system. What follows is organized in four main sections. Our summarizer benefits from a number of linguistic analysis filters; these, as well as some ....

Marcu, Daniel. 1997. From discourse structures to text summaries. In Proceedings of ACL'97 Workshop on Intelligent, Scalable Text Summarisation, pages 82--88, Madrid, Spain.

.... as a whole (or of some coherent segment of the document) A variety of approaches fall into this general category, ranging from fairly common sentence extraction techniques to newer methods utilising, for example, strong notions of topicality [4] 8] lexical chains [3] and discourse structure [14], 5] see the papers from the recent ACL workshop on Intelligent, Scalable Text Summarization [2] for relevant overview) Ultimately, all of these approaches share a fundamental similarity: they construct a characterisation of document content through significant reduction of the original ....

D. Marcu. From discourse structures to text summaries. In Proceedings of ACL'97 Workshop on Intelligent, Scalable Text Summarisation, pages 82--88, Madrid, Spain, 1997.

....approach (Kupiec et al. 1995) uses a corpus of articles with summaries for training to identify the features of sentences that are typically included in abstracts. Other approaches use lexical chains (Barzilay and Elhadad 1997) sentence position (Lin and Hovy 1997) discourse structure (Marcu 1997; Marcu 1998) and user features from the query (Strzalkowski et al. 1998) to find key sentences. While most of the work to date focuses on summarization of single articles, early work is beginning to emerge on summarization across multiple documents. Radev and McKeown 1998) use a symbolic ....

Daniel Marcu. From Discourse Structures to Text Summaries. In Proceedings of the ACL Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, August 1997. Association for Computational Linguistics.

.... more sophisticated techniques are being deployed in attempts to improve the quality of sentencebased summaries, by seeking to mediate the passage selection process with, for instance, strong notions of topicality (Hovy Lin 1997) lexical chains (Barzilay Elhadad 1997) and discourse structure (Marcu 1997), Reimer Hahn 1997) 1 Also at: http: www.nytimes.com library cyber digicom 012797digicom.html. 1.3 Capsule overviews The approach we take in this work, while addressing a slightly different problem to that of strict summarisation, can be construed as striving for the best of both worlds. ....

Marcu, D. 1997. From discourse structures to text summaries. In Proceedings of ACL'97 Workshop on Intelligent, Scalable Text Summarisation, 82--88.

....problem through the seventies and eighties (e.g. 17, 25] The resources devoted to addressing this problem grew by several orders of magnitude with the advent of the world wide web and large scale search engines. Several innovative approaches began to be explored: linguistic approaches (e.g. [1, 2, 4, 12,14,15, 18]) statistical and information centric approaches (e.g. 6, 9, 16, 24] and combinations of the two (e.g. 3, 24,26] The TIPSTER Phase III Program, an information retrieval initiative of the US Defense Department funded several of these projects on summarization [27] Almost all of this work ....

Marcu, D. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization (Madrid, Spain, 1997), pp. 82--88.

....approach [Kupiec et al. 1995] uses a corpus of articles with summaries for training to identify the features of sentences that are typically included in abstracts. Other recent approaches use lexical chains [Barzilay and Elhadad 1997] sentence position [Lin and Hovy 1997] discourse structure [Marcu 1997; Marcu 1998] and user features from the query [Strzalkowski et al. 1998] to find key sentences. While most work to date focuses on summarization of single articles, early work is emerging on summarization across multiple documents. Radev and McKeown [1998] use a symbolic approach, pairing ....

Daniel Marcu. From Discourse Structures to Text Summaries. In Proceedings of the ACL Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, August 1997. Association for Computational Linguistics.

....By making use of the CSTI corpus, we were able to reduce the problem of generating summary discourse to the problem of generating a sequence of such message types and by ne tuning them to produce coherent discourse using realization switches. Unlike schemas [McKeown, 1985] or RST [Hovy, 1988, Marcu, 1997, Moore and Paris, 1989] the multi sentence generation process in summons is guided primarily by the chronological order of the templates and only rarely requires more than one sentence type per input template. Two distinctions exist between the texts in the CSTI corpus and the summaries ....

....how various discourse processing techniques (e.g. rhetorical structure relations) can be used to both identify important information and form the actual summary. While promising, this work does not involve an implementation as of yet, but provides a framework and strategies for future work. [Marcu, 1997] uses a rhetorical parser to build rhetorical structure trees for arbitrary texts and produces a summary by extracting sentences that span the major rhetorical nodes of the tree. 138 In addition to domain speci c information extraction systems, there has also been a large body of work on ....

Daniel Marcu. From discourse structures to text summaries. In Proceedings of the Workshop on Intelligent Scalable Text Summarization, pages 82-88, Madrid, Spain, July 1997. Association for Computational Linguistics.

....defines the macrolevel semantic structure of a connected discourse, wh2 ile cohesion creates connectedness in a non structural manner. Coherence is represented in terms of coherence relations between text segments, such as elaboration, cause and explanation. Some researchers, e.g. 23] and [19], use discourse structure (encoded using RST [18] as a source representation for summarization) Discource representation can be used to prune a hierarchical tree of discourse segments and keep only the nucleus of the discourse. In contrast to lexical cohesion, however, coherence is difficult to ....

....from distant places in the text for a single chain. The complete results of our experiments are available on line at http: www.cs.bgu.ac.il summarization test. 4 Evaluation of the Method 4. 1 Description of the Experiment Most evaluations of summarization systems use an intrinsic method [4, 24, 16, 19, 27, 23]. The typical approach is to create an ideal summary, either written by professional abstractors or by merging summaries provided by multiple human subjects using methods such as majority opinion, union, or intersection. The output of the summarizers is then compared with the ideal summary. ....

[Article contains additional citation context not shown here]

Daniel Marcu. From discourse structures to text summaries. In ACL/EACL-97 summarization workshop, pages 82--88, Madrid, 1997.

....of the source; iii) the condensation of the extracted information and construction of a summary representation; and (iv) the presentation of the summary representation to the reader in natural language. While some techniques exist for producing summaries for domain independent texts (Luhn, 1958; Marcu, 1997) it seems that domain specific texts require domain specific techniques (DeJong, 1982; Paice and Jones, 1993) In our case, we are dealing with technical articles which are the result of the complex process of scientific inquiry that starts with the identification of a knowledge problem and ....

.... and with others published with source documents (usually author abstracts) We have chosen Microsoft 97 Summarizer because, even if it only produces extracts, it was the only summarizer available in order to carry out this evaluation and because it has already been used in other evaluations (Marcu, 1997; Barzilay and Elhadad, 1997) In order to evaluate content, we presented judges with randomly selected abstracts and five lists of keywords (content indicators) The judges had to decide to which list of keywords the abstract belongs given that different lists share some keywords and that they ....

D. Marcu. 1997. From Discourse Structures to Text Summaries. In The Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82-- 88, Madrid, Spain, July 11.

....as essential in the process of abstracting. 1 Introduction The idea of producing abstracts or summaries by automatic means is not new, several methodologies have been proposed and tested for automatic abstracting including among others: word distribution (Luhn, 1958) rhetorical analysis (Marcu, 1997); and probabilistic models (Kupiec et al. 1995) Even though some approaches produce acceptable abstracts for specific tasks, it is generally agreed that the problem of coherent selection and expression of information in automatic abstracting remains (Johnson, 1995) One of the main problems is ....

D. Marcu. 1997. From discourse structures to text summaries. In The Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, July 11.

....By making use of the CSTI corpus, we were able to reduce the problem of generating summary discourse to the problem of generating a sequence of such message types and by fine tuning them to produce coherent discourse using realization switches. Unlike schemas [McKeown, 1985] or RST [Hovy, 1988, Marcu, 1997, Moore and Paris, 1989] the multi sentence generation process in summons is guided primarily by the chronological order of the templates and only rarely requires more than one sentence type per input template. Two distinctions exist between the texts in the CSTI corpus and the summaries ....

....how various discourse processing techniques (e.g. rhetorical structure relations) can be used to both identify important information and form the actual summary. While promising, this work does not involve an implementation as of yet, but provides a framework and strategies for future work. [Marcu, 1997] uses a rhetorical parser to build rhetorical structure trees for arbitrary texts and produces a summary by extracting sentences that span the major rhetorical nodes of the tree. 138 In addition to domain specific information extraction systems, there has also been a large body of work on ....

Daniel Marcu. From discourse structures to text summaries. In Proceedings of the Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, July 1997. Association for Computational Linguistics.

....868; 227 web pages, each annotated with a short (roughly 13 word) human authored summary. Some important prior work in extractive summarization has explored issues such as cue phrases [13] positional indicators [6] lexical occurrence statistics [15] and the use of implicit discourse structure [14]. Most of this work relies fundamentally on a property of the source text which web pages often lack: a coherent stream of text with a logical discourse structure. Somewhat closer in spirit to ocelot is work on combining an information extraction phase followed by generation; for instance, the ....

Marcu, D. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization (1997), pp. 82--88.

....The information used for automated score prediction by e rater canalsobeusedas building blocks for automated generation of diagnostic and instructional summaries. Clauses and sentences annotated by APA as the beginning of a new argument might be used to identify main points of an essay (Marcu (1997)) In turn, identifying the main points in the text of an essay could be used to generate feedback reflecting essay topic and organization. Other features could be used to automatically generate statements that explicate the basis on which e rater generates scores. Such statements could supplement ....

Marcu, Daniel. (1997). "From Discourse Structures to Text Summaries.", In Proceedings of the Intelligent Scalable Text Summarization Workshop, Association for Computational Linguistics, Universidad Nacional de Educacion a Distancia, Madrid, Spain.

....that can be used to improve the results of discourse based summarizers. Previous work review of the experiment In the experiment on discourse trees and summaries (Marcu 1997a; 1997c) I used five texts from Scientific American that ranged in size from 161 to 725 words. I broke the texts into 160 elementary textual units and asked 13 independent judges to rate each textual unit according to its importance to a potential summary. The judges assigned a score of 2 to the units that they believed to be very important, 1 to those that they considered somewhat important, and 0 to those that ....

Text 8(3):243--281. Marcu, D. 1997a. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, 82--88.

....the results. In section 3, we describe the experiment on task based evaluation methods. The last section is conclusion. Ideal Summary Based Evaluation Previous Work Most evaluations of summarization systems use an intrinsic method (Edmunson 1969; Paice 1990; Kupiec, Pederson, Chen 1995; Marcu 1997; Salton et al. 1997; Ono, Sumita, Miike 1994) The typical approach is to create an ideal summary, either by professional abstractors or by merging summaries provided by multiple human subjects using methods such as majority opinion, union, or intersection. The output of the summarizers is ....

....quite consistent with respect to what they perceive as being the most important and the most unimportant but less consistent with respect to what they perceive as being less important. The percent agreement in our experiment is surprisingly high compared to results presented by other researchers. Marcu (Marcu 1997) found percent agreement of 13 judges over 5 texts from Scientific America is 71 . Rath (Rath, Resnick, Savage 1961) found that extracts selected by four different human judges had only 25 overlap. Salton (Salton et al. 1997) found that the most important 20 paragraphs extracted by 2 subjects ....

[Article contains additional citation context not shown here]

Marcu, D. 1997. From discourse structures to text summaries. In ACL/EACL-97 summarization workshop, 82--88.

....problem through the seventies and eighties (e.g. 29, 37] The resources devoted to addressing this problem grew by several orders of magnitude with the advent of the world wide web and large scale search engines. Several innovative approaches began to be explored: linguistic approaches (e.g. [3, 4, 8, 19, 24, 25, 30]) statistical and information centric approaches (e.g. 10,16,20,27,36] and combinations of the two (e.g. 6, 36, 38] The TIPSTER Phase III Program, an information retrieval initiative of the US Defense Department funded several of these projects on summarization [39] Human quality ....

D. Marcu. From discourse structures to text summaries. [1], pages 82--88.

....problem through the seventies and eighties (e.g. 19, 26] The resources devoted to addressing this problem grew by several orders of magnitude with the advent of the worldwide web and large scale search engines. Several innovative approaches began to be explored: linguistic approaches (e.g. [2, 3, 6, 12, 15, 16, 18, 20]) statistical and informationcentric approaches (e.g. 8, 9, 17, 25] and combinations of the two (e.g. 5, 25] Almost all of this work (with the exception of [12,16,20,24] focused on summarization by text span extraction , with sentences as the most common type of text span. This ....

Marcu, D. From discourse structures to text summaries. [1], pp. 82--88.

....1964] possible negative factors (for instance, words that might indicate lesser significance) Mathis et al. 1973] etc. More recently, Salton and his colleagues experimented with probabilistic measures for word importance [Salton et al. 1997] Marcu looked at learning structural importance [Marcu, 1997] , and Hovy and Lin looked at machine learning approaches for positional importance [Hovy and Lin, 1997] In contrast to the large amount of work that has been undertaken in extractive summarization, there has been much less work on a generative model of summarization. The earliest approaches to ....

Daniel Marcu. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, 1997.

....c 1999, American Association of Artificial Intelligence (www.aaai.org) All rights reserved. structure (Marcu 1997) combinations of information extraction and language generation (Klavans Shaw 1995; McKeown, Robin, Kukich 1995) and the use of machine learning to find patterns in text (Teufel Moens 1997; Barzilay Elhadad 1997; Strzalkowski, Wang, Wise 1998) However, it is difficult to compare the relative merits of these various approaches because most of the evaluations reported were conducted on different corpora, of varying sizes at varying levels of compression, and were often informal ....

Text 8(3):243--281. Marcu, D. 1997. From discourse structures to text summaries.

.... structure of the input text and its description in terms of rhetorical relations in order to eliminate the information units of secondary importance and condense the content of the whole text in a top down fashion (Mani and Maybury, 1997) Recently, some progress has been made in the (e.g. (Marcu, 1997), Scheler and Fischer, 1997) the automatic identification of rhetorical relations and other types of pragmatic information in terms of surface cues. The resulting environments are largely domain independent, because they are based on the text structuring devices and their surface expression ....

D. Marcu. From Discourse Structures to Text Summaries. In I. Mani and M. Maybury, editors, Proceedings of the ACL/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, 11 July, UNED, Madrid, Spain, 1997.

.... summarization through rhetorical parsing tuning Daniel Marcu Information Sciences Institute University of Southern California 4676 Admiralty Way, Suite 1001 Marina del Rey, CA 90292 6601 marcu isi.edu Abstract We study the relationship between the structure of discourse and a set of summarization heuristics that are employed by current systems. A ....

Daniel Marcu. 1997a. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, July 11.

No context found.

Daniel Marcu. From Discourse Structures to Text Summaries. In Proceedings of the Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, August 1997. Association for Computational Linguistics.

No context found.

Marcu D. From discourse structures to text summaries. Proc of the Workshop on Intelligent Scalable Text Summarization, Association for Computational Linguistics 1997: 82-88.

No context found.

D. Marcu. From Discourse Structures to Text Summaries. The Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pp. 82-88, 1997.

No context found.

D. Marcu. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, 1997.

No context found.

D. Marcu. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, 1997.

No context found.

D. Marcu. From discourse structure to text summaries. In Proceedings of ACL/EACL-97 summarization workshop, pages 82-88, 1997.

No context found.

D. Marcu, "From discourse structures to text summaries," presented at 14th National Conference on Artificial Intelligence (AAAI-97), 1997.

No context found.

Daniel Marcu. From discourse structure to text summaries. In Proceedings of the ACL '97 EACL '97 Workshop on Intelligent Scalable Text Summarization, pages 82-88, Madrid, Spain, July 1997.

No context found.

Daniel Marcu. From Discourse Structures to Text Summaries. In The Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, July 11 1997.

No context found.

Daniel Marcu. 1997. From discourse structures to text summaries. In Proceedings of the ACL Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, August.

No context found.

Daniel Marcu. 1997. From discourse structure to text summaries. In ACL/EACL-97 Workshop on Intelligent and Scalable Text Summarization.

No context found.

Daniel Marcu. 1997. From discourse structures to text summaries. In Proceedings of the ACL Workshop on Intelligent Scalable Text Summarization, pages 82-88, Madrid, Spain, August. Association for Computational Linguistics.

No context found.

Daniel Marcu. 1997. From discourse structures to text summaries. In Inderjeet Mani and Mark T. Maybury, editors, Proceedings of the workshop on Intelligent Scalable Text Summarization, in association with ACL/EACL-97.

No context found.

Text 8(3):243--281. Marcu, D. 1997. From discourse structures to text summaries. In ACL/EACL-97 Workshop on Intelligent Scalable Text Summarization, 82--88. Madrid: Association for Computational Linguistics and the European Chapter of the Association for Computational Linguistics.

No context found.

Marcu, D. (1997a): From discourse structures to text summaries. 82-88 in Mani, I.; Maybury, M. eds.: Intelligent Scalable Text Summarization. Workshop Sponsored by the Association for Computational Linguistics. Madrid.

No context found.

"From discourse structures to text summaries" in Working Notes of the ACL'97 Workshop on Intelligent Scalable Text Summarization

No context found.

"From discourse structures to text summaries" in Working Notes of the ACL'97 Workshop on Intelligent Scalable Text Summarization

No context found.

Daniel Marcu. 1997. From discourse structures to text summaries. In Proceedings of the ACL Workshop on Intelligent Scalable Text Summarization, pages 82--88, Madrid, Spain, August.

No context found.

Marcu, D. 1997. From discourse structures to text summaries. In Mani, I. and Maybury, M. eds. Intelligent scaleable text summarisation, 82-88. Proceedings of a Workshop Sponsored by the ACL. Somerset NJ: Association for Computational Linguistics.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC