| { |
| "paper_id": "J75-2002", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T02:39:57.510740Z" |
| }, |
| "title": "", |
| "authors": [ |
| { |
| "first": "Ja", |
| "middle": [], |
| "last": "Chauche", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Nathematiques Appliquees -Informatique Universite", |
| "location": {} |
| }, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "ATEF converts an input string into a labeled tree; t h e label evolves under the c o n t r o l of a grammar. A s e t of labels is associated with each segment of the string, and several functions permit control of the number of alternative labels. CETA simulates a transformatianal grammar. It uses a set o f grammars with conditional linkages. The applicabili&y of a transformation can be determined in p a r t by conditions on the resulting t r e e. Computer processing 3f natural languages requires more or less polished algorithmic models. The two systems presented here represent a choice of a large class among the algorithms proposed in recent years to solve these problems. The princ i p a l choice determined by these systems l i e s in the formal use of labeled trees (arborescences). Freedom of choice of these labels and possible structures gives these systems broad fields", |
| "pdf_parse": { |
| "paper_id": "J75-2002", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "ATEF converts an input string into a labeled tree; t h e label evolves under the c o n t r o l of a grammar. A s e t of labels is associated with each segment of the string, and several functions permit control of the number of alternative labels. CETA simulates a transformatianal grammar. It uses a set o f grammars with conditional linkages. The applicabili&y of a transformation can be determined in p a r t by conditions on the resulting t r e e. Computer processing 3f natural languages requires more or less polished algorithmic models. The two systems presented here represent a choice of a large class among the algorithms proposed in recent years to solve these problems. The princ i p a l choice determined by these systems l i e s in the formal use of labeled trees (arborescences). Freedom of choice of these labels and possible structures gives these systems broad fields", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "of applications in several domains and notably in that of the automatic processing of natural languages. The ATEF system has the purpose of transformFng a string of words into a tree which is manipulable by the CETA system. The definition of labeled trees determines what objects CETA can manipulate and the objecti'ves of ATEF. This note therefore begins with the definition of labeled trees. To obtain a tree of this type beginning with an input string, ArEF uses a dictionary and a finite-state grammar. The result of this system can be manipulated by CETA in order to obtain the desired type of structure. T h e example of analysis given here shows t h e possibilities of the CETA system w?th two different manipulative strategies: search for constituent or dependency structure .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "A rree is a set of points with which is associated a structure, that is to say a relation having the properties:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1, LABELED TREES", |
| "sec_num": null |
| }, |
| { |
| "text": "The relation between two points is directed (one point depends on the o t h e r )", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1, LABELED TREES", |
| "sec_num": null |
| }, |
| { |
| "text": "A point Cannot depend on a point belonging to its own descent set (the descent set of a point is the set of points that depend on it, the points that depend on them, etc. )", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1, LABELED TREES", |
| "sec_num": null |
| }, |
| { |
| "text": "A unique point descends from no other.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1, LABELED TREES", |
| "sec_num": null |
| }, |
| { |
| "text": "It is possible to draw a tree placing below a point all of its descendants, linked by lines. The definitdon of a particular label consists in an enumeration of the variables relevant to t h e label. A s e t of labels can be predefined and is collected in a so-called format file.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1, LABELED TREES", |
| "sec_num": null |
| }, |
| { |
| "text": "A dictionary is a set of segments (character strings), ~t h each of which is associated a label, a processing pointer, and a lexical unit pointer. The processing pointer specifies the particular process which must be associated with the segment.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| }, |
| { |
| "text": "T h e a n a l y s i s of t h e input word by the ATEF system resides at first in a label processing, that is to say in a n evolution of the empty label toward a final label characteristic of the analyzed word This evolution is controlled by the grammar, which at each moment has access to two labels. the label being developed (noted by the symbol C) and t h e label associated w i t h the segment which was read in the dictionary (noted by A)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| }, |
| { |
| "text": "The analysis of a word aims to produce a segmentation of t h e woru simultaneously compatible with the segments of t h e d i f f e r e n t dictionaries (the word must be an assembly of dictionary segments) and compatible with a correct evolution of the grammar The choice and evolution of the segmentation has to do with the sequence of input characters. The segmentation forces, above all, a prior linguistic choice. Thus with the segment \"UN\" two possibilities can be conceived either accept \"UN E\" as a coherent segmentation or have the segment \"Ul?E\" in the dictionary and refuse the segmentation \"UN El'", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| }, |
| { |
| "text": "For each initial form several segmentations are possible to arrive at the same results arid only a linguie tic study of the phenomena permits a decision on the strategy to be adopted.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| }, |
| { |
| "text": "In any event, this stLategy is left to the user of the system", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| }, |
| { |
| "text": "In the course of a segmentation the system can operate d~rectly on the nonsegmented chatacters in order to force them into a \"canonical'! form. Thus in the case of the word reel several possib~lities arise to accept a word like realite put the segment \"real\" in a dictionary as well as the segmenf \"reel\", the former will generate words like realite, irrealite, e t c .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| }, |
| { |
| "text": "put the single segment \"reel\" in the dictionary and the analysis of the word realite Qill follow the schema realit& => 1st segment found \"lte\", remainder \"real\" mddificat&on r e a l ->r&el => 2nd segment found \"reel\" segmentation .\"riSel it&\" N B In thls analysis, it is to be noted that the search for successive segqents is performed from left to right \u00a3ox the input word. This depends on the strategy adopted and, for a given use, the direction of the segmentatson of a word can be either left to right or right to l e f t . (Observe that the segmentation of the word \"chacune\" will then be o b t a i n e d as G A C UNE because th8 segmentation CHACUN E will. be rejected as a subsegmentation of \"UNE\" This problem can easily be resolved because t h e s e functions appear in t h e rules of the grammar and are consequently conditional. One can at t h e same time f\"orbid the subsegmentation \"UN E-\" i n the word \"UNE\" and aur;horize this. segmentatian in the word \"CHACUNE\")", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| }, |
| { |
| "text": "The calculation of the set of labels associated ~i t h a word is produced and controled by the grammar. This calculation corresponds above all with a conditianal modification of the label C or current state starting from the label A or argument state. rejected. This condition can refer to the labels of the preceding analyzed words and can condition its result on the analysis of the following form. Thus for example in the course of the analysis of the word \"LA\" in the sequence \"il la voit\", the segmentation taking \"la\" as a r t i c l e can be rejected. The t r a n s - ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| }, |
| { |
| "text": "fer", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The ATEF system analyzes the words and thus employs dictionaries", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": { |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Arborescence e t transformation. Thesis, Grenoble", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Chauche", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Chauche, J . Arborescence e t transformation. Thesis, Grenoble. December 19.74", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Translation of a t c x~ prepared for the First National Gonference on Computational Linguistics", |
| "authors": [], |
| "year": 1975, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Translation of a t c x~ prepared for the First National Gonference on Computational Linguistics, Varna, May, 1975.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF0": { |
| "text": "(See the example on the next frame. ) A labeled tree is a tree such that with each of its points is associated a lab-el This l a b e l IS formed of a set of data. The figure below represents a labeled tree. fernthe,-nom commurr feminin, sin&ul.the ATEF system is to transform an input sming of words i n t o a labeled t r e e , each word in the string p o s s i b l y , l e a d i n g to one or several points in the final tree (ambiguity). The determination of the label originating in an input word results from its analysis. This analysis proceeds by segme~tation of the i n p u t word according t o elements f r o m d i f f e r e n t dictional: ies . A correct segmentation therefore gives. a l a b e l for a p o i n t of t h e f i n a l tree . Tn advance of any. analysis, the definition of the elements employed in the composition of different labels is required and is supplied by two Eiles called variable declaration files. A label will consist of a set bf variables. Each variable must be defined with its set of possible values. Thus if one defines the variable \"category\" the set of \"categories\" that can be used must be specified. The set is w r i t t e n category = (NOUN, ARTICLE, PRONOUN, ADJECTIVE, VERB, . . . ) (A constraint requires that the name of a variable must not be longer than 7 characters. Thus the preceding var2able could be written, for example, CAT = (NN, ART, Pm, ADJ, VRB, . . . ) )", |
| "type_str": "figure", |
| "num": null, |
| "uris": null |
| }, |
| "FIGREF1": { |
| "text": "the segmentation of the input ~o r d is tighcly bound to the evolution of the grammar which controls the coherence of the segmentation In the course of a segmentation operation the state of the system takes into account for the analyzed word t h e l a b e l resuLting from the analysis of t h e segments a l~e a d y obtarned for this word the label associated w i t h the segment found in a dictlonary the remaining characters of the input word the complete form of the i n p u t word Thus for example in the course of the analysis of the word irresistible and after analysis of the segment \" i b l e \" and in the course of reading the segment 'resist1' the following erements are obtained C the label resulting from the analysis of \"ible\" This label contains for example the v a r i a b l e derivation with value verb-adj, the variable gender with value masculine and feminine, the variable number with value singular. A the label associated w i t h the segment''resist T h i s label coatains notably the lexical unit \"resister\", the varlable category with value verb The character$ IR The complete form IRRESISTIBLE The purpose of the grammar is to permit or prevent the evolution of label C starting wlth label A Here, the l a b e l will evolve and obtaln the variable category wlth value adjective. A rule associated with the segment \"resist\" by means of its pointer will therefore describe this evolution of the label C. When no evolution of the label C is possible, the corresponding segmentation is blocked and considered nonsignificant. The set of labels plays a fundamental role in this system and forms the s e t of s t s~t e s o f t h e f i n i t e state transducer corresponding to t h e l o g i c a l model of the system Each coherent segmentation of a word (a word can have several coherent segmentations leading to ambiguities) provides a labeled point in the final tree Three elements are fundamental to the system the choice and evolution of the segmentation the calculation of the set of labels associated with a word the positioning of the labeled points created by the analysis of a word in the final tree", |
| "type_str": "figure", |
| "num": null, |
| "uris": null |
| }, |
| "FIGREF2": { |
| "text": "To avoid a proliferation of possible segmentations and therefore of possible solutions, several functions provide for intervention in the segmentation A first possibility is offered b :~ the management o f the dictionaries In fact, the system includes several dictionaries and after isolation of each segment the I I system can open\" or \"close\" a dictionary This method makes it easy to avoid, for example, looking \u20acor two consecutive prefixes. Another mode of intervention uhich is more d i r e c t , i s peovided by the presence of functions acting on the enumeration procedures by which the system counts ~f f solutions. For example, t h e system analyzes a l l possible segmentations starting with a given segment beginning with the segmentations containing most characters. An intervention at this level makes i t possible h o t t o analyze but to reject subsegmentations of P segment. The analysis of the segment \"UNE\"can, for example, r e j e c t the analysis of t h e subsegment \"UN E\"", |
| "type_str": "figure", |
| "num": null, |
| "uris": null |
| }, |
| "FIGREF3": { |
| "text": "condition f o r t h e evolution of t h i s label is such that if no evolution is possible then the corresponding segmentation Ls", |
| "type_str": "figure", |
| "num": null, |
| "uris": null |
| }, |
| "FIGREF4": { |
| "text": "of information to d i f f e r e n t labels can be realized through a s s i g n m e n t t o t h e following label S . When this label has been assigned i n the course of t h e analysis of a word t h e analysis of the follow'ing word w i l l begin with the assigned label instead ef the null l a b e l . The final result of the system i s a labeled tree. With no supplementary specification i n the course o f analysis, this tree appears in the following form: The solution for a sentence (~h r a s e ) consists of a string of l a b e l s (one f o r each word of t h e sentence), each of which represents an interpretation of a ward of t h i s sentence. In this c a s e , t h e sentence is not structured; simply the ambiguities are s e p a~a t e d . Th t h e course of t h e a n a l y s i s of the words, a first sketch of a construct5an can be made and give as r e s u l t a m o r e developed tree. These functions s p e c i f y the p o s i t i o n that the point t o which t h e calculated mask applies must take in the final tree. This position is determined in a l l cases below a p o i n t w, and is relative to t h e root ( f i r s t p o i n t on the l e f t below wi) and to the rightmost p o i n t of the tree a l r e a d y conn a l y s i s of une\", no t r e e during the analysis of \"belle\", the tree eontains the single point \"une\". A function can render the point \"belle\" as root and give b e l l e une during the analysis of \"maison\", if the constructed tree is belle,. a function can provide for swapping w e t h e root with t h e occurrence i n work and give the tree une belle. In this case, the result f o r the system w i l l be une belle 5 , THE CETA SYSTEM The CETA system provides for writing and s5mulating a transformational grammar. This system manipulates l a b e l e d trees of the type described above ( l a b e l e d trees produced by the ATEF or other system). To construct a transformational grammar with this system two complementary elements are necessary: the set of rules u s e d d e f i n e s the s e t of prfmitives of the system for a given application The set of grammars and the d e f i n i t i o n of t h e i r linkage defines the mode of use of the primitives The definition of a transformation rule defines a mode of potential transformation of the tree considered. A rule is defined by a le'fthand part representing the subtree to be modif i e d and a righthand part d e f i n i n g the resulting subtree. For example, l e t t h e following be two rules: r e s u l t i n g f r o m t h e analysis b y t h e ATEF system of t h e senterice \"une i r r e s i s t i b l e jeune femme rousse\", we w i l l have t h e following applications: une i r r g s l s t i b l e j e u r r e femme rousse irresistible . i b u n z GT*' r o u s s e e s i s t i b l e j e u n e femme 'ousse i r r 6 s i s t i b l e J e u n e femme r o u s s e art. a d j . ad j ..In fact, the~defin'itton of a transformation can call on ahierarchical set of subtrees. In t h e example taken here, the i n p u t tree is not very \"deep\" and most often only one-level trees are applicable. However, in t h e course of development of a cornplete structure, tshe considered tree is arbitrary and the definition of a complex transformation constructed beginning with several subtrees is very refined. The subtrees defined in a rule can likewise be considered.", |
| "type_str": "figure", |
| "num": null, |
| "uris": null |
| }, |
| "FIGREF5": { |
| "text": "rules makes. it possible t o obtain complex constructions by simulating r e p e t i t i v e procedures. A recursive rule is charaoterized by a call to a new grammar (which can obviously be the same as thae in which the recursive rule is found). The result o f ' t h e application of a recursive rule consists of the tree obtained after transformation by the called grammar of rhe t r e e transformed by the rule in question. For example, let R3 and R2 be the-rules previously described. The elementary grammar G consistiag of these two rules will furnish as result, In unitary mode: application of R3. (Priority is given by the order of enumeration of t h e r u l e s . ) Grammar G' containing rules R3 and R 2 , but specifying t h a t R3 L I~U S~ be recursive and call G', we have: a CETA grammar consT-s-t-s of a set of elementary grammars and a conditional linking procedure over them. The linking must he such t h a t t h e Corresponding graph is loop-free. An elementary gxamar from which no linking is possible yields as result t h e input tree i n place of the transformed tree. This procedure permits one to obtain a methad of analysis involving several c r i t e r i a of acceptance, each consisting i n t h e presence of a t r e e schema i n t h eterminal tree.", |
| "type_str": "figure", |
| "num": null, |
| "uris": null |
| }, |
| "TABREF0": { |
| "content": "<table><tr><td>J</td><td>1</td><td>j</td><td>An</td></tr><tr><td colspan=\"4\">elementary grammar has furthermore a mode of execution. Atl ele-</td></tr><tr><td colspan=\"4\">mentary grammmar unitarily executable is such that its result</td></tr><tr><td>will be obtained</td><td/><td/><td/></tr></table>", |
| "type_str": "table", |
| "num": null, |
| "html": null, |
| "text": "The linkage of the different rules previously described is defined by the set of elementary grammars. An elementary grammar c~~n~i s t s of ordered rules. A rule Ri will be applied prior to an R. if the order of R . is less ahan the order of R after an application of a part of the set of rules mentioned. (An application of the rules mentioned can cause to appear new possible applications which will not be performed in this case.) Anoth-er mode of application of an elementary grammar is exhaustive. In this mode, the set of rules of the gzammar will be applied up ro the maximum but the application of a given rule has the effect o f eliminating it from this elementary grammar. (That is, for a given point.) With this second mode of application, the number of possible steps for a given tree is always finite. Within an elementary grammar which is unitarily or exhaustively exechtable, the presence of recursive" |
| } |
| } |
| } |
| } |
