| { | |
| "paper_id": "W98-0132", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T06:03:48.145005Z" | |
| }, | |
| "title": "Memoisation in Sentence Generation with Lexicalised Grammars", | |
| "authors": [ | |
| { | |
| "first": "Nicolas", | |
| "middle": [], | |
| "last": "Nicolov", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "This paper discusses a sentence generation system PROTECTOR which uses: (i) a non-hierarchical semantic representation which allows for flexible lexical choice and un~form treatment of different languages, (ii) a lexicalised D-Tree Grammar which is very similar to Tree-Adjoining Grammar in spirit, and {iii) dynamic programming techniques to avoid doing redundant computations. We review the motivation for choosing such an organisation of the generalion syslem and give an example of the generation of a sentence which involves a lexical gap. The generation of the example sentence requires a nondeterministic mode of computation (the lexical gap forcing backtracking). We show how dynamic programming techniques can be used to save regenerating structures using a top-down generation algorithm.", | |
| "pdf_parse": { | |
| "paper_id": "W98-0132", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "This paper discusses a sentence generation system PROTECTOR which uses: (i) a non-hierarchical semantic representation which allows for flexible lexical choice and un~form treatment of different languages, (ii) a lexicalised D-Tree Grammar which is very similar to Tree-Adjoining Grammar in spirit, and {iii) dynamic programming techniques to avoid doing redundant computations. We review the motivation for choosing such an organisation of the generalion syslem and give an example of the generation of a sentence which involves a lexical gap. The generation of the example sentence requires a nondeterministic mode of computation (the lexical gap forcing backtracking). We show how dynamic programming techniques can be used to save regenerating structures using a top-down generation algorithm.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "Natural language generation is the process of generating text from a set of abstract communicative goals. lt attempts to model the human language production mechanisms in rnanmachine communication. As part of the overall generation process computer systems will need to consider how the communicative goals can be mapped onto conceptual representations and these in turn into sentences in a natural language. The latter process is known as sentence generation and this paper discusses a system for doing this task (realising sentences from meaning representations).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Early work on sentence generation assumed input of the form: pred(arg 1 \u2022 \u2022\u2022\u2022 argn) and the generaiion process was reduced to mapping prad --+ verb, arg 1 --+ first complement, etc. This approach, of course, makes the 11 semantic structures\" be nothing more than disguised syntactic representations and reduces the sentence generation problern to finding out the ordering of the constituents. The tree-like semantic assurnption does not allow for handling head switching examples (Nicolov, 1993) , incorporation of rnodifiers in the syntactic head ( French blond and blond French girl cannot be generated from franch(blond(girl))) and cases like: She smiled a welcome to the guests. / She welcomed the guests with a smile.", | |
| "cite_spans": [ | |
| { | |
| "start": 480, | |
| "end": 495, | |
| "text": "(Nicolov, 1993)", | |
| "ref_id": "BIBREF12" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conceptual input", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "Such phenomena can be addressed more elegantly using non-hierarchical semantic representations. In PROTECTOR conceptual graphs are used (Sowa, 1992) . The same generation mechanisms can be used with underspecified discourse representation structures.", | |
| "cite_spans": [ | |
| { | |
| "start": 136, | |
| "end": 148, | |
| "text": "(Sowa, 1992)", | |
| "ref_id": "BIBREF17" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conceptual input", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "D-Tree Grammar (Rambow et al., 1995) is a grammar formalism which arises from work on Tree-Adjoining Grammars (TAG) (Joshi, 1987) . 1 In the context of generation, TAGS have been used in a number of systems MUMBLE (Mc-Donald and Pustejovsky, 1985) , SPOKESMAN (Meteer, 1990) , WIP (Wahlster et al., 1991) , synchronous TAGS (Shieber and Schabes, 1991) the system reported by McCoy (McCoy et al., 1992) , the first version of PROTECTOR (Nicolov et al., 1995) , and SPUD (Stone and Doran, 1997) . TAGS have been given a prominent place in the VERBMOBIL project -they have been chosen to be the framework for the generation module (Caspari and Schmid, 1994; Harbusch et al., 1994; Becker et al., 1998) . In the area of grammar development TAG has been the basis of one of the largest grarnmars developed for English (Doran et ai., i994) . 1 DTG and TAG are very similar, yet they are not equivalent (Weir pc). Unlike TAGs, DTGs provide a uniform treatment of complementation and modification at the syntactic level. DTGs are seen as attractive for generation because a close match between semantic and syntactic operations leads to simplifications in the overall generation architecture. DTGs try to overcome the problems associated with TAGS while remaining faithful to what is seen as the key advantages of TAGS (Joshi, 1987) :", | |
| "cite_spans": [ | |
| { | |
| "start": 15, | |
| "end": 36, | |
| "text": "(Rambow et al., 1995)", | |
| "ref_id": "BIBREF14" | |
| }, | |
| { | |
| "start": 116, | |
| "end": 129, | |
| "text": "(Joshi, 1987)", | |
| "ref_id": "BIBREF6" | |
| }, | |
| { | |
| "start": 132, | |
| "end": 133, | |
| "text": "1", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 214, | |
| "end": 247, | |
| "text": "(Mc-Donald and Pustejovsky, 1985)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 250, | |
| "end": 274, | |
| "text": "SPOKESMAN (Meteer, 1990)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 277, | |
| "end": 304, | |
| "text": "WIP (Wahlster et al., 1991)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 324, | |
| "end": 351, | |
| "text": "(Shieber and Schabes, 1991)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 375, | |
| "end": 401, | |
| "text": "McCoy (McCoy et al., 1992)", | |
| "ref_id": "BIBREF8" | |
| }, | |
| { | |
| "start": 435, | |
| "end": 457, | |
| "text": "(Nicolov et al., 1995)", | |
| "ref_id": "BIBREF11" | |
| }, | |
| { | |
| "start": 469, | |
| "end": 492, | |
| "text": "(Stone and Doran, 1997)", | |
| "ref_id": "BIBREF18" | |
| }, | |
| { | |
| "start": 628, | |
| "end": 654, | |
| "text": "(Caspari and Schmid, 1994;", | |
| "ref_id": "BIBREF1" | |
| }, | |
| { | |
| "start": 655, | |
| "end": 677, | |
| "text": "Harbusch et al., 1994;", | |
| "ref_id": "BIBREF4" | |
| }, | |
| { | |
| "start": 678, | |
| "end": 698, | |
| "text": "Becker et al., 1998)", | |
| "ref_id": "BIBREF0" | |
| }, | |
| { | |
| "start": 813, | |
| "end": 827, | |
| "text": "(Doran et ai.,", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 828, | |
| "end": 833, | |
| "text": "i994)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 1311, | |
| "end": 1324, | |
| "text": "(Joshi, 1987)", | |
| "ref_id": "BIBREF6" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "D-Tree Grammars", | |
| "sec_num": "3" | |
| }, | |
| { | |
| "text": "1. the extended domain of locality over which syntactic dependencies are stated; and 2. function argument structure is captured within a single initial construction in the grammar. 4 Generation strategy PROTECTOR uses declarative specification of the relation between semantics and syntax encoded as mapping rules. The mapping rules are elementary d-trees (i.e\u00bb tree descriptions) annotated with applicability semantics a match with which will licence the applicability of the mapping rule. In addition if the d-tree has nonterminal leaf nodes relevant parts of the applicability semantics are related to these nodes so that we know how the sernantics is decornposed. PROTECTOR employs a top-down (recursive descent) strategy for generating the complements once an initial top-level mapping rule lias been chosen (this stage is called generation of skeletal structure). PROTECTOR keeps track liow rnuch of the input semantics it has consurned. Then in a consequent stage the remaining semantics is consumed which involves the use of modification and sister-adjunction.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "D-Tree Grammars", | |
| "sec_num": "3" | |
| }, | |
| { | |
| "text": "In this section we discusss the generation of a sentence which involves a lexical gap:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "*Alexander attacked the town 'fu/1-scalely'. Alexander launched a full-scale attack on the town.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "The input semantics and the search space are are shown in Figure 3 (see next page). At the onset of generation there are at least two toplevel mapping rules that can be chosen ( attack and launch an attack) and the default one ( attack) leads to a dead end. The reason is the lack of a mapping rule (not only in the linguistic knowledge base of the generator but worse of all in the English language) that would allow us to express the concept 1FULL-SCALE1 as a structure that we can intergrate to the existing skeletal syntactic structure (Alexander attacked the town). Such is the nature of lexical gaps and this forces backtracking. The generator would need to reconsider its previous decisions, it would have to undo (forget) about all the structures it had built all the way up to the point when it chose the wrang mapping rule. This was the first choice that was made so practically cvcry computa.tion is lost. i\\.11", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 58, | |
| "end": 66, | |
| "text": "Figure 3", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Example", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "the work that went into building the subject and object NPs has to be duplicated. Figure 3 : The search space for the example and generating its required complements will result in re-computation of the subject and object NPs. These NPs can be arbitrarily large and in order to avoid doing redundant computations we store the results of previous generation goals and reuse them if needed again. Such dynamic programming techniques have been exploited heavily in parsing and PROTECTOR's declarative mapping rules and flexibility of incorporat-ing alternative generation strategies allows us to take advangates of that work. This approach is gaining popularity in generation (Shieber, 1988; Haruno et al., 1993; Pianesi, 1993; Gerdemann and Hinrichs, 1995; Kay, 1996; Nicolov et al., 1997) . The other approaches to chart generation are based on CFGs and in a bottom-up strategy one has to make sure that in moving from an N to NP all modifiers have been ex-pressed. This causes serious overhead in backtracking. Our use of DTGs and flexible way of adding modifiers using precedence constraints between semantic classes of modifiers does not suffer from this problem.", | |
| "cite_spans": [ | |
| { | |
| "start": 673, | |
| "end": 688, | |
| "text": "(Shieber, 1988;", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 689, | |
| "end": 709, | |
| "text": "Haruno et al., 1993;", | |
| "ref_id": "BIBREF5" | |
| }, | |
| { | |
| "start": 710, | |
| "end": 724, | |
| "text": "Pianesi, 1993;", | |
| "ref_id": "BIBREF13" | |
| }, | |
| { | |
| "start": 725, | |
| "end": 754, | |
| "text": "Gerdemann and Hinrichs, 1995;", | |
| "ref_id": "BIBREF3" | |
| }, | |
| { | |
| "start": 755, | |
| "end": 765, | |
| "text": "Kay, 1996;", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 766, | |
| "end": 787, | |
| "text": "Nicolov et al., 1997)", | |
| "ref_id": "BIBREF11" | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 82, | |
| "end": 90, | |
| "text": "Figure 3", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Example", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "PROTECTOR does not assume that lexical choice is performed prior to surface realisation.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "lt chunks the input semantics appropriately on the basis of the mapping rules.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Example", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "We have described a sentence generator which takes non-hierarchical input, uses mapping rules to relate parts of the semantics to elementary d-trees, combines the syntactic structures in a manner that closely mirrors the semantic decomposition and employs dynamic programming to avoid re-generation of structures on backtracking which cannot always be predicted in advance as is the case for lexical gaps. Our architecture allows for easy encoding of alternative generation strategies (e.g., bottom-up, best-:first, etc.) which other systems have not considered and in fact find rather difficult to do. Thus, PROTECTOR can be seen as a test bed for experimenting and evaluating alternatives methods for generation.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusions", | |
| "sec_num": "6" | |
| } | |
| ], | |
| "back_matter": [], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "An efficient kerne! for multilingual generation in speech-to-speech die.logue transle.tion", | |
| "authors": [ | |
| { | |
| "first": "Tilman", | |
| "middle": [], | |
| "last": "Becker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Wolfgang", | |
| "middle": [], | |
| "last": "Finkler", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Anne", | |
| "middle": [], | |
| "last": "Kilger", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Peter", | |
| "middle": [], | |
| "last": "Poller", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1998, | |
| "venue": "Procudings of COLING-AGL'98)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Tilman Becker, Wolfgang Finkler, Anne Kilger, and Peter Poller. 1998. An efficient kerne! for multilingual genera- tion in speech-to-speech die.logue transle.tion. In Procud- ings of COLING-AGL'98), Montreal, Canada.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Parsing und generierung in trug", | |
| "authors": [ | |
| { | |
| "first": "Rudolf", | |
| "middle": [], | |
| "last": "Caspari", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ludwig", | |
| "middle": [], | |
| "last": "Schmid", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1994, | |
| "venue": "", | |
| "volume": "40", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Rudolf Caspari and Ludwig Schmid. 1994. Parsing und generierung in trug. Technical Report Verbmobil-Report 40, Siemens AG, December.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "XTAG -A Wide Coverage Grammarfor English", | |
| "authors": [ | |
| { | |
| "first": "Christine", | |
| "middle": [], | |
| "last": "Doran", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Dania", | |
| "middle": [], | |
| "last": "Egedi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Beth", | |
| "middle": [ | |
| "Ann" | |
| ], | |
| "last": "Hockey", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Bangalore", | |
| "middle": [], | |
| "last": "Srinivas", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Martin", | |
| "middle": [], | |
| "last": "Zaidel", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1994, | |
| "venue": "Proceedings oj the 15th Int. Conference on Computational Linguistics (COLING-94)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "5--9", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Christine Doran, Dania Egedi, Beth Ann Hockey, Bangalore Srinivas, and Martin Zaidel. 1994. XTAG -A Wide Cov- erage Grammarfor English. In Proceedings oj the 15th Int. Conference on Computational Linguistics (COLING-94), pages 922-928, Kyoto, Japan, 5-9 August.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Some open problems in head-drivcn generation", | |
| "authors": [ | |
| { | |
| "first": "Dale", | |
| "middle": [], | |
| "last": "Gerdemann", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Erhard", | |
| "middle": [], | |
| "last": "Hinrichs", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "Linguistics fJ Computation", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "65--197", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Dale Gerdemann and Erhard Hinrichs. 1995. Some open problems in head-drivcn generation. In Jerry Morgan, Georgia Green, and Jenn.ifer Cole, editors, Linguistics fJ Computation, pages 65-197. CSLI Publications.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Default handling in incremental generation", | |
| "authors": [ | |
| { | |
| "first": "G", | |
| "middle": [], | |
| "last": "Harbusch", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Kilger", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1994, | |
| "venue": "Proceedings oj the 15th International Conference on Computational Linguistics (COLING-94)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "l<. Harbusch, G. ]{ikui, and A. Kilger. 1994. Default han- dling in incremental generation. In Proceedings oj the 15th International Conference on Computational Linguis- tics (COLING-94).", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Bidirectional chart generation of natural language texts", | |
| "authors": [ | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Masahiko", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yasuharu", | |
| "middle": [], | |
| "last": "Haruno", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yuji", | |
| "middle": [], | |
| "last": "Den", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Makoto", | |
| "middle": [], | |
| "last": "Matsumoto", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Nagao", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1993, | |
| "venue": "Proceedings of the I Ith National Conference on Artijicial lntel/igence (AAAI'93)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "350--356", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Masahiko. Haruno, Yasuharu Den, Yuji Matsumoto, and Makoto Nagao. 1993. Bidirectional chart generation of natural language texts. In Proceedings of the I Ith Na- tional Conference on Artijicial lntel/igence (AAAI'93), pages 350-356, Menlo Park, CA. American Association for Artificiai lntelligencc, The MIT Press.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "The Relevance ofTree Adjoirnng Grammar to Generation", | |
| "authors": [ | |
| { | |
| "first": "Aravind", | |
| "middle": [], | |
| "last": "Joshi", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1987, | |
| "venue": "Na.tural Language Generation", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "233--252", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Aravind Joshi. 1987. The Relevance ofTree Adjoirnng Gram- mar to Generation. In Gerard Kempen, editor, Na.tural Language Generation, pages 233-252. I<luwer Academic, Dordrecht, The Netherlands.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "Chart generation", | |
| "authors": [ | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Martin L\\ay", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1996, | |
| "venue": "Proceedings of the S4th Ann1Ja/ Afeeting oj the Auociation Jor Computational Linquistics (ACL '96 )", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "200--204", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Martin l\\ay. 1996. Chart generation. In Proceedings of the S4th Ann1Ja/ Afeeting oj the Auociation Jor Computational Linquistics (ACL '96 ), pages 200-204, Santa Cruz, Califor- nil'o.", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "A functional approach to generation with tag", | |
| "authors": [ | |
| { | |
| "first": "F", | |
| "middle": [], | |
| "last": "I<athleen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Mccoy", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Gijoo", | |
| "middle": [], | |
| "last": "Vijay-Shanker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Yang", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Proceedings oj the SOth A1mual Meeting of the Associatiot1 Jor Computational Linguistics(ACL '92)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "48--55", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "I<athleen F, McCoy, K. Vijay-Shanker, and Gijoo Yang. 1992. A functional approach to generation with tag. In Proceed- ings oj the SOth A1mual Meeting of the Associatiot1 Jor Computational Linguistics(ACL '92), pages 48-55, Univer- sity of Delaware, U.S. Association of Computational Lin- guistics.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "TAGs as a grammatical formalism for generation", | |
| "authors": [ | |
| { | |
| "first": "David", | |
| "middle": [], | |
| "last": "Mcdonald", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "James", | |
| "middle": [], | |
| "last": "Pustejovsky", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1985, | |
| "venue": "Proceedings of the 2.'lrd Annual Meeting oj the Association Jor Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "94--103", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "David McDonald and James Pustejovsky. 1985. TAGs as a grammatical formalism for generation. In Proceedings of the 2.'lrd Annual Meeting oj the Association Jor Computa- tional Linguistics, pages 94-103.", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "The \"Generation Gap", | |
| "authors": [ | |
| { | |
| "first": "Marie", | |
| "middle": [], | |
| "last": "Wenzel Meteer", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1990, | |
| "venue": "The Problem of E,;preuibility in Te,;t Planning", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "90--94", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Marie Wenzel Meteer. 1990. The \"Generation Gap\": The Problem of E,;preuibility in Te,;t Planning. Ph.D. thesis, Computer and Information Science Department, Univer- sity of Massachusetts, February. COINS Technical Report 90-04.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "Approximate Chart Generation from Non-Hierarchical Rcpresentations", | |
| "authors": [ | |
| { | |
| "first": "Nicolas", | |
| "middle": [], | |
| "last": "Nicolov", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Chris", | |
| "middle": [], | |
| "last": "Mellish", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Graeme", | |
| "middle": [], | |
| "last": "Ritchie", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "Ruslan Mitkov & Nicolas Nicolov, editor, Recent Advances in Natural Language Processing: Selected papers Jrom RANLP '95, Current lssues in Linguistic Theory (CILT)", | |
| "volume": "136", | |
| "issue": "", | |
| "pages": "273--294", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Nicolas Nicolov, Chris Mellish, and Graeme Ritchie. 1995. Sentence Generation from Conceptual Graphs. In Gerard Ellis, Robert Levinson, William Rich, and John Sowa, edi- tors, LNAI 954, Conceptua/ Structures: Applications, Im- plementation and Theory, pages 74-88. Springer, Berlin, 14-18 August. Proceedings of the 3rd Int. Conf. on Con- ceptual Structures (ICCS'95), Santa Cruz, CA, USA. Nicolas Nicolov, Chris Mellish, and Graeme Ritchie. 1997. Approximate Chart Generation from Non-Hierarchical Rcpresentations. In Ruslan Mitkov & Nicolas Nicolov, edi- tor, Recent Advances in Natural Language Processing: Se- lected papers Jrom RANLP '95, Current lssues in Linguistic Theory (CILT), 136, pages 273-294. John Benjamins, Am- sterdam & Philadelphia.", | |
| "links": null | |
| }, | |
| "BIBREF12": { | |
| "ref_id": "b12", | |
| "title": "Head Selection in NLG. DAI discussion paper 140", | |
| "authors": [ | |
| { | |
| "first": "Nicola", | |
| "middle": [], | |
| "last": "Nicolov", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1993, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Nicola.s Nicolov. 1993. Head Selection in NLG. DAI discus- sion paper 140, Dept. of Artificial Itclligence, Univ. of Ed- inburgh.", | |
| "links": null | |
| }, | |
| "BIBREF13": { | |
| "ref_id": "b13", | |
| "title": "Head-driven bottom-up generation and Govemment and Binding: a unified perspectivc", | |
| "authors": [ | |
| { | |
| "first": "Fabio", | |
| "middle": [], | |
| "last": "Pianesi", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1993, | |
| "venue": "New Concepts in Natural Language Generation", | |
| "volume": "3", | |
| "issue": "", | |
| "pages": "187--214", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Fabio Pianesi. 1993. Head-driven bottom-up generation and Govemment and Binding: a unified perspectivc. In Hel- mut Horacek and Michael Zock, editors, New Concepts in Natural Language Generation, chapter 3, pages 187-214. Pinter, London.", | |
| "links": null | |
| }, | |
| "BIBREF14": { | |
| "ref_id": "b14", | |
| "title": "D-Tree Grammars", | |
| "authors": [ | |
| { | |
| "first": "Owen", | |
| "middle": [], | |
| "last": "Rambow", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Vijay-Shanker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "David", | |
| "middle": [], | |
| "last": "Weir", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "Proceedings oj the 33rd Meeting of the Aaaociation Jor Computational Linguistics (ACL'95)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "151--158", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Owen Rambow, K. Vijay-Shanker, and David Weir. 1995. D-Tree Grammars. In Proceedings oj the 33rd Meeting of the Aaaociation Jor Computational Linguistics (ACL'95), pages 151-158.", | |
| "links": null | |
| }, | |
| "BIBREF15": { | |
| "ref_id": "b15", | |
| "title": "Generation and synchronous trec-adjoing grammars", | |
| "authors": [ | |
| { | |
| "first": "Stuart", | |
| "middle": [], | |
| "last": "Shieberand", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yves", | |
| "middle": [], | |
| "last": "Schabes", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1991, | |
| "venue": "Special issue on Natural Language Generation", | |
| "volume": "7", | |
| "issue": "4", | |
| "pages": "220--228", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Stuart Shieberand Yves Schabes. 1991. Generation and syn- chronous trec-adjoing grammars. Computational Intelli- gence, 7( 4):220-228. Special issue on Natural Language Generation.", | |
| "links": null | |
| }, | |
| "BIBREF16": { | |
| "ref_id": "b16", | |
| "title": "A uniform archltecture for parsing and generation", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Stuart", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Shleber", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1988, | |
| "venue": "COLING-88", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Stuart M. Shleber. 1988. A uniform archltecture for parsing and generation. In COLING-88, Budapest.", | |
| "links": null | |
| }, | |
| "BIBREF17": { | |
| "ref_id": "b17", | |
| "title": "Conceptual graphs summary", | |
| "authors": [ | |
| { | |
| "first": "F", | |
| "middle": [], | |
| "last": "John", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Sowa", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Conceptua/\u2022Structures: Current Reaearch and Practice", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "3--51", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "John F. Sowa. 1992. Conceptual graphs summary. In Tim- othy E. Nagle, Janice A. Nagle, Laurie L. Gerholz, and Peter W. Eklund, editors, Conceptua/\u2022Structures: Current Reaearch and Practice, Ellis Horwood Series in Workshops, pages 3-51. Ellis Horwood Limited, London, England.", | |
| "links": null | |
| }, | |
| "BIBREF18": { | |
| "ref_id": "b18", | |
| "title": "Sentence Planning as Description Using Tree Adjoining Grarnmar", | |
| "authors": [ | |
| { | |
| "first": "Matthew", | |
| "middle": [], | |
| "last": "Stone", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Christine", | |
| "middle": [], | |
| "last": "Doran", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1997, | |
| "venue": "Procetdinga oj the 35th A nnual Afeeting oj the Association of Computo.tional Linguistics(ACL'97)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "198--205", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Matthew Stone and Christine Doran. 1997. Sentence Plan- ning as Description Using Tree Adjoining Grarnmar. In Procetdinga oj the 35th A nnual Afeeting oj the Associa- tion of Computo.tional Linguistics(ACL'97), pages 198- 205, Madrid, Spain.", | |
| "links": null | |
| }, | |
| "BIBREF19": { | |
| "ref_id": "b19", | |
| "title": "WIP: the coordinated generation of multi modal presentations from a common representation", | |
| "authors": [], | |
| "year": 1991, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Wolfgang Wahlster, Elisabeth Andre Son Bandyopadl.Jya)', Winfried Graf, and Thomas Rist. 1991. WIP: the coordi- nated generation of multi modal presentations from a com- mon representation. RR 91-08, DFKI.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "text": "operations to combine elementary structures -subsertion (Figure 1) and sister adjunction (Figure 2). The elementary structures are d-trees ( descriptions of trees) which in addition to immediate dominance relation allow for stating dominance relationships between nodes in the d-tree.", | |
| "num": null, | |
| "uris": null, | |
| "type_str": "figure" | |
| }, | |
| "FIGREF1": { | |
| "text": "e use a lexicalised ( every elementary structure contains a terminal node (anchor) which 'justifies' the construction), feature-based (nonterminals are feature structures) DTG.", | |
| "num": null, | |
| "uris": null, | |
| "type_str": "figure" | |
| } | |
| } | |
| } | |
| } |