File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/metho/83/e83-1029_metho.xml
Size: 25,423 bytes
Last Modified: 2025-10-06 14:11:35
<?xml version="1.0" standalone="yes"?> <Paper uid="E83-1029"> <Title>NATURAL LANGUAGE INPUT FOR SCENE GENERATION M</Title> <Section position="4" start_page="0" end_page="175" type="metho"> <SectionTitle> I CLAUSE \[.- </SectionTitle> <Paragraph position="0"> ,.1 ........ ...... ............</Paragraph> <Paragraph position="1"> h ! v--.o.</Paragraph> <Paragraph position="2"> arrows indicate reading from input. Each</Paragraph> <Paragraph position="4"> level is activated by the superior one, as shown in Fig.l, and returns to its caller the results of its computation as a message. A feature network is associated to each process, which is activated together with its corresponding processes.</Paragraph> <Paragraph position="5"> In a PASCAL-like language the feature network can be defined as follows: type FEATURE .... (.LIST OF FEATURES.)...;</Paragraph> <Paragraph position="7"> Each NODE represents s feature identified by its NAME; the ALTERNATE pointer allows the connection in a Circular list of mutually exclusive features as in SHRLDU (Winograd,72). Each process gives as output a fragment of the FEATURE NETWORK manipulated to describe the input; this is performed by means of a set of functions which test the presence of a feature in the FEATURE_NETWORK, add and erase features, as described in McCord ('77). The process is divided into a set of sequential routines,called SLOTs, analyzing the functional components of a Syntactic Unit. In the function: function FILLER(ARGI:PROCESS, ARG2:SETOF_FEATURES):boolean; ARGI activates the appropriate process to fill the caller slot; the second argument of the function selects the set of features to which the called process must be inizialized.</Paragraph> <Paragraph position="8"> This last features-passing mechanism is absent in the original PSG; from our experience, we found it usefull in all the cases in which a choice in a syntactic level is determined by the syperior level or by a more larger context.</Paragraph> <Paragraph position="9"> Thus, for instance, the set of features characterizing a prepositional phrase is determined at the corresponding syntactic level by the preposition and the features of the nominal phrase; but further and not less important selection criteria can be imposed by the verb which is found in the upper level. The output of a simple analysis is shown in Fig.2; it gives an idea of the syntactic representation. null</Paragraph> </Section> <Section position="5" start_page="175" end_page="175" type="metho"> <SectionTitle> INPUT: IL M~O GATTO STA MANGIANDO {my ca~ is eating) RESULT OF THE ANALYSIS: TIPO DICHIARAT.TVA ATTIVA (dec lara~ive active} IL MIO GATTO SOGGETTO {suOJeC/~} STA MANG IANDO VERB . GR FEATURE NETWORK : </SectionTitle> <Paragraph position="0"> --CLAUSE- ! -PROPOSIZIONE- ! -PRINC- ! -DICHIARATIVA--I -VERB. GR ...... ! -PE&quot;S-- ! -TERZA-! I -NUM --- ! -SINGOLARE-! ! -MOUO-- ! -ESPLICITO- ! - INDICATIVO--I ! -TEMPO- ! -PRESENTE--I ! -ACT--- f -TRANSITIVA-! -FORMA- ! -STARE-! f -GENERICO-! -COMPL. UR ..... ! -PERS-- ! -TERZA-! -GEN--- ! -MASCHILE--gUM--- ~-SZNGOLARZ--NOUN.C~-~-.ONE-- null f -TIPO-I-COMUNE PSRS--FIG.2 - Result of a Sentence Analysis The choice of PSG is mainly motivated by the possibility of parallel computation. A control structure allowing the parallel computation is: cobegin ... coend; It is a single input-output structure, very usefull to handle alternative choices for the same computational level. In the case of mutually exclusive alternatives only one of the &quot;n&quot; processes activated by a cobegin control structure can end successfully. In the case of not mutually exclusive alternatives, it is still possible to use the cobegin control structure , but it is necessary to define a strategy for the selection of the most suitable alternative when the coend occurs.</Paragraph> <Paragraph position="1"> An experimental implementation in terms of para~ lel computation has been made on a multiprocessor system (Adorni et ai.,'79). Another version of this parser has been implemented in PASCAL (DiManzo et ai.,'79} and a version in FranzLisp is in progress.</Paragraph> </Section> <Section position="6" start_page="175" end_page="175" type="metho"> <SectionTitle> III STRUCTURE OF THE COGNITIVE DATA BASE </SectionTitle> <Paragraph position="0"> The organization of knowledge, in this system, is based on a set of THOUGHTs. A THOUGHT is a frame like structure within which new data are interpreted in terms of concepts acquired through previous experience (Minsky,'75), (Schank,Abelson,'77).</Paragraph> <Paragraph position="1"> Every THOUGHT has a TYPE which determines a set of operations applicable to it. The following predefined types are allowed (Adorni,DiManzo,'83): - DESCRIPTIVE, that defines the complete description of a physical,abstract,animate or not,object. - PROTOTYPE, that defines the structural part of a physical object in terms of generalized cones (Marr,Nishihara,'78). An example of definition of simple prototype object is given in Fig.3.</Paragraph> <Paragraph position="2"> - JOINT, that defines the element of connection between physical objects, in order to build more complex objects or scenes (Fig.4).</Paragraph> <Paragraph position="3"> - SPATIALREL, that defines spatial relationships like &quot;on,near,on the left of,...&quot; between objects. All the linguistic relationships like &quot;above,under, behind&quot;, and so on, are reduced into quantitative geometrical relationships between the coordinates of some points of the involved objects; this choice is motivated by the possibility of deriving a set of very general inference rules from analytic geometry (Adorni et ai.,'82), (Boggess,'79), (Boggess, Waltz,'79). The coordinates of an indefinite point P are given in the form:</Paragraph> </Section> <Section position="7" start_page="175" end_page="177" type="metho"> <SectionTitle> COORD K OF P (REFERRED_TO A)=H </SectionTitle> <Paragraph position="0"> where K is a group of possible coordinates, H a set of values for these coordinates and A is the THOUGHT of the object to which the reference system used is connected. Fig.5 shows the THOUGHT for an use of the preposition &quot;on&quot;.</Paragraph> <Paragraph position="1"> A spatialrel type THOUGHT can contain conceptualizations and prototype THOUGHTs; a joint type can contain only its description; a prototype type can contain joint or prototype THOUGHTs or descriptions in terms of generalyzed cones;all these types can be enclosed in a descriptive type which can contain conceptualizations and all the types of THOUGHTs, previously introduced. A descriptive type can include the following fields (Adorni,DiManzo, '83), (see Fig.6): - DESCR, contains all the basic not specialized knowledge about the object; - LEVELS, contains a description of the shape of the object (in terms of prototype THOUGHTs) divided in different levels of detail hierarchically organized; - USE, contains the descriptions of the most common activities involving the use of the object, in terms of spatialrel between prototype THOUGHTs; FIG.3 - Example of Definition of a Simple Prototype FIG.4 - Definition of a Simple Jointing Element and Use of this Element to build a More Complex Object - POSITION, gives the most common spatial relations between the described object and other objects in standard scenes, in terms of a spatialrel between prototype THOUGHTs; - SUPPORT, contains the indication, in terms of descriptive THOUGHTs, of the objects which are supported in standard situations; - COLOR and MADE, describe the possible set of colors and materials, while WEIGHT contains information about the range of possible weights; - CONTENT, says, in terms of descriptive THOUGHTs, that the normal use of the object is a container for other objects; - DYNAMIC, contains the current expectations about the boundaries of the dimensions of the objects; it can be dinamically updated every time a new object of the same class enters the system's CDB.</Paragraph> </Section> <Section position="8" start_page="177" end_page="179" type="metho"> <SectionTitle> IV SEMANTIC INTERPRETER </SectionTitle> <Paragraph position="0"> The Semantic Interpreter of the IA interacts with the Syntactic Analyzer and operates on a set of rules in order to build the concepts a sentence was intended to mean. The output of this module is a Conceptual Dependency Network (Schank,'75), in which every nominal is substituted by a complex descriptive THOUGHT instantiated from the CDB.</Paragraph> <Paragraph position="1"> Let us illustrate the procedure of analysis considering the following sentence (the translation is word by word in order to reproduce the problems of Italian): (i) &quot;l'uomo dai capelli grigi e' andato a Roma con l'auto di Giuseppe&quot; (the man with the grey hair has gone to Rome with the car of Joseph) The procedure of analysis has several steps: A. Analysis of Words and Simple Phrases During this step the entities which take part into the conceptualization are identified. In fact an indexed identifier Xi is associated to each object referred to in the sentence (each nominal), which points to one or more conceptualizations, contained in the field &quot;descr&quot; of each nominal in the CDB. The adjectives contained in the noun phrases are also analyzed during this step. Each of them adds some conceptualizations which contribute to further individuate the nominal. During this step personal pronouns are identified as: Xi ~=--> ISA(HUMAN) Temporal and local adverbials are also analyzed in this phase in order to assign to the sentence conceptualization a time and place identification according to certain rules described in (Adorni et al.,'81).</Paragraph> <Paragraph position="2"> At the end of this step the sentence (i) is represented as follows: identifier nominal conceptualization</Paragraph> <Paragraph position="4"> The sentence (i) can then be read: (2) &quot;XI da X2 e' andato a X3 con X4 di X5&quot; (XI from X2 is gone to X3 with X4 of X5) B. Analysis of Modifiers The simple phrases of a sentence can either fill conceptual cases of a main conceptualization, thus serving as 'picture producer' (PP), or further ind !</Paragraph> <Paragraph position="6"> {exists, at least, ~ poin? P which is part Of the THOUGhT(A) and a point ~ which is par= of r.he THOUGHT(B) and for a.ny paL.&quot; Op points P and 0 is Z(P) >. ~(Q). More, there is an tssertion about the fact that the THOUGHT(B) suppor~ the THOUGHT(A)).</Paragraph> <Paragraph position="7"> viduate a PP. Therefore they can be classified according to whether they modify: a) the nominal that precedes(also not immediately); &quot;i libri di Carlo&quot; ^ (the books of Charles) b) the subject or object independently from their position; &quot;Maria e' andata a Roma con Anna&quot; ^ (Mary has gone to Rome with Ann) c) the action; &quot;Maria e' andata a Roma con la macchina&quot; ^ (Mary has gone to Rome with the car) thought ... end. . ) NF.A~ IS $vatlalrel t~ought ... end. ) FIG.6 - Definition of the Descriptive THOUGHT</Paragraph> <Section position="1" start_page="178" end_page="179" type="sub_section"> <SectionTitle> of a Chair </SectionTitle> <Paragraph position="0"> The treatement of the modifiers in b) and 4) requires that the structure of the sentence is entirely known and cannot, in any case, be performed before the verb has been analyzed (subject and object are considered type c) modifiers). The modifiers in a), on the contrary, have a local role, limited to the PP they are to modify, and their relation to the sentence structure is marginal.</Paragraph> <Paragraph position="1"> They are, therefore,immediately associated to their corresponding nominals. In (2) &quot;da X2&quot; and &quot;di X5&quot; are of this kind and are consequently linked to X1 and X4 producing: (3) &quot;XI e' andato a X3 con X4&quot; (XI has gone to X3 with X4) In the &quot;descr&quot; field of THOUGHTs Xl and X4 the following information is added:</Paragraph> <Paragraph position="3"> The embodying of a modifier creates complex PPs or CLUSTERs. Each CLUSTER has as its HEAD a b) or c) modifier,a conceptual index node modified by the accessory concepts.</Paragraph> <Paragraph position="4"> In our example &quot;l'uomo dai capelli neri&quot;, &quot;a Roma&quot;, and &quot;con l'auto di Giuseppe&quot; are CLUSTERs, in which the head is always the leftmost nominal.</Paragraph> <Paragraph position="5"> The decision about the embodying of a modifier into its head is related to the classical problem of the placement of PP's. In fact, it is not always the case that a prepositional phrase modifies a conceptual index node; it is often possible that it has to be embodied into another accessory modifier, as in: &quot;il libro dell'uomo dal cappotto blu&quot; (the book of the man with the blue coat) If it is defined: md ---> the current phrase; md-i ---> the immediately proceeding phrase; md-2 ---~ the phrase that immediaZely proceeds md-l; the solution is obtained by recursively deciding wether md is to be embodied into md-i or md-2. Recursion is from the lower level.</Paragraph> <Paragraph position="6"> This decision is made by a set of standard general procedures associated to prepositions (di, da, con, per .... ) and adverbs (sopra, sotto, davanti, diefro, ...). Non-standard specialized alternatives are activated by particular nouns and verbs in order to treat hidiosyncrasies. These procedures are written as three-steps programs, which accomplish the operations of: 1-LOOKING for compatibility of certain features of md,md-l, and md-2. Typical features are superset and part-of relations of md's. A rule may state that &quot;IF md has a part-of relation to md-2 THEN md may be embodied into md-2&quot;. Example: &quot;il libro del bambino dalla copertina rossa&quot; /---md-2 ...... md-I ........... md ............ / (the book of the child with red cover) 2-Deciding whether MERGING can be performed. This is made by imposing further restrictions of the type described above. Also the main conceptualization and other linguistic peculiarities are taken into account.</Paragraph> <Paragraph position="7"> 3-Actual LINKING. In our example, the structure: md <===> PART OF(md-i OR md-2) &quot;l'uomo dai capelli grigi&quot; /--md-I ...... md .......... / HAIR <===> PART_OF(MAN) is produced because md &quot;capelli&quot; can be part-of md &quot;uomo&quot;.</Paragraph> <Paragraph position="8"> Should it not have been the case, the following structure would have been produced: (md-I OR md-2) < .... POSS(md) &quot;l'uomo dal vestito scuro&quot; /--md-i ....... md ......... / C. Construction of the Main Conceptualization The nucleus of a main conceptualization can be associated in the CDB both to a noun indicating an action, state or change of state and a verb. In our example, we find the THOUGHT of fig. 8. A time identification always related to the present (T@), is taken from the syntactic analysis and connected to this conceptualization, thus resulting into:</Paragraph> <Paragraph position="10"> If a lexical ambiguity arises, the features assumed by the nominals in the previous steps will help to desambiguate.</Paragraph> <Paragraph position="12"> At this step &quot;splitting&quot; of a conceptualization often occurs. In the sentence: &quot;Giovanni d~ un colpo a Maria&quot; (lit. John gives a blow to Mary) although two nuclei are present (d~ & colpo),nevertheless the correct interpretation is &quot;Giovanni colpisce Maria&quot; (John hits Mary), instead of &quot;Giovanni trasferisce il possesso dell'oggetto colpo a Maria&quot; (John tansfers the ownership of the object 'blow' to Mary)!!! We have observed that this phenomenon involves conceptualizations based on the primitives of &quot;state&quot;, &quot;action&quot;, and &quot;spatial relationship&quot; and relies only on the pairs ACTION-STATE, ACTION-SPATIAL RE-LATIONSHIP, and ACTION-ACTION. The regularities ruling the formation of these pairs have been found to depend only upon those conceptual primitives.</Paragraph> <Paragraph position="13"> This keeps the number of rules to be evaluated reasonably small, if compared with the number of CDB entries (~600 entries in the present implementation (Adorni et al.,'81))~ An example will illustrate the mechanism of reduction of the conceptual &quot;splitting&quot; as well as of disambiguation.</Paragraph> <Paragraph position="14"> The pair ACTION-SPATIAL RELATIONSHIP may be represented by: &quot;tirare su il braccio&quot; ^ A</Paragraph> </Section> </Section> <Section position="9" start_page="179" end_page="180" type="metho"> <SectionTitle> ACTION SPATIAL RELATIONSHIP </SectionTitle> <Paragraph position="0"> (lift the arm) The compound &quot;tirare su&quot; has the two meanings: - innalzare, alzare,...(lift,raise .... ); - confortare, dare sollievo psiehico,...(encourage, console .... ); which can be conceptualized respectively: X ..... PTRANS .... OBJ(Y) .... DIR ( (FROM (K))(TO(H) ) ) and ((COORD Z OF H- COORD Z OF K) and R(X ..... PROPEL .... OBJ(Y) .... DIR((FROM(NIL)) The context helps disambiguation. In our example, the object of the spatial relationship being a physical object, the first alternative is selected. The rule performs a further control, discovering that the physical object is, in this case, PART OF(HUMAN); the PROPEL primitive is then substituted by the MOVE primitive. D. Case Fillin~ in the Main Conceptualization The next step performed by the semantic module is the filling of the conceptual cases of the main conceptualization with the THOUGHTs instantiated during the previous steps. Again, standard rules are associated to prepositions and adverbs and hidiosyncrasies are also treated. These rules make use of messages sent by the syntactic component and look at the conceptual syntax of the main conceptualization. Through these rules the cluster&quot;con X4&quot; turns out to be 'instrumental' and the following conceptualization is then produced: (4) X1 .... USE .... OBJ(X4) Since the filler of the instrumental case of the main conceptualization has to be a conceptualization, the rule activated by the &quot;con&quot; modifier fills the instrumental case with (4).</Paragraph> <Paragraph position="1"> In (3), 'a X3' is placed in the destination of the directive case of the main conceptualization, because preposition 'a' is stated to indicate the 'destination' if the main conceptualization contains a PTRANS,PROPEL or MOVE,with empty directive case; otherwise it indicates 'state'. &quot;Andare a Roma&quot; is thus distinguished from &quot;essere a Roma&quot; (to be in Rome). The result, for our example, is:</Paragraph> <Paragraph position="3"> The directive case,as shown in the above example is not simply filled with a md; it is filled with a &quot;spatial_relationship-md&quot; pair. This is a general rule for our system, emphasizing the change of coot dinates caused by an action. In our example this means that the primitive PTRANS has moved the object to a point whose coordinates are defined with in the city of Rome. The result of the analysis of (I) is given in Fig.9.</Paragraph> <Paragraph position="5"> E. Conceptual Analysis of Complex Sentences The process of semantic interpretation is applied to every clause in the sentence, identified by a verb or a noun indicating an action. Segmentation into such clauses or nominalized clauses is obviously performed by the syntactic component, which has also non-standard rules for specific classes of (modal) verbs like: dovere (must),volere (to want),potere (can),incominciare (to start) ....</Paragraph> <Paragraph position="6"> These verbs constitute a single main conceptualization together with the embedded infinitive. Simple composition rules have been defined to combine the meaning of clauses (sentences).</Paragraph> <Paragraph position="7"> Thus for conjunction, as in &quot;si alzo',si mise il cappello eapri' la porta&quot; (he stood up,put on its hat and opened the door) the main conceptualizations associated to every proposition are connected by an 'and' relationship.</Paragraph> <Paragraph position="8"> (si alzo') ......................... T1 and (si mise il cappello) T2 >TI and (apri' la porta) T3 >T2 A time indication is also associated to every main conceptualization to emphasize the execution order of every action.</Paragraph> <Paragraph position="9"> Conceptual analysis of each single clause (sentence) is activated by this top level structure and at the end the resulting conceptualizations are linked one to the other.</Paragraph> </Section> <Section position="10" start_page="180" end_page="181" type="metho"> <SectionTitle> V CONCLUSIONS </SectionTitle> <Paragraph position="0"> In this paper a system for understanding a natural language input to a scene generator has been described. It makes use of a conceptual dependency semantic model, substantially modified in as much as syntax is kept apart from semantic interpretation and a fully formalized dictionary is used, much more complex than the one embodied in Schank's theory. The dictionary is particularly oriented to the generation of scenes, and the stress is on the representation of the structure of objects.</Paragraph> <Paragraph position="1"> The awareness of the structure of the objects is often intimately related to our capability of under standing the. meaning of spatial relationships and other complex linguistic expressions. For instance, the meaning &quot;the cat is under the car&quot; is clear, even if it may depend on the state of the car, moving or parked; on the contrary, the sentence &quot;the cat is under the wall&quot; is not clear, unless the wall is crashed or it has a very particular shape.Our model tries to account t~is understanding activity by means of the following features: - an object is described at several levels of details; in some cases, only a rough definition of the object dimensions can be sufficient, while in other cases a more sophisticated knowledge about the structure of the object itself is required; null - the characteristic features of an object are emphasized; the recognition of a feature allows the activation of particular rules and the generation of hypotheses about the presence of an object; null - the typical relationships among objects are described.</Paragraph> <Paragraph position="2"> The interaction between syntactic and semantic analyzers seems rather complex, but it provides some valuable solutions to certain crucial points of computational linguistics, like PP's placement, conceptual splitting, idioms and preassembled The syntactic analyzer, working top-down, yelds a representation of the input sentence in which information about gender, number, person and tense are recorded and for each function such as subj, obj, time, etc.., the ccrresponding filler is identified, or a list of fillers is given in case of ambiguity. These two kinds of information are exactly what is usefull for semantic interpretation and are picked up in various steps of the interaction by the semantic analyzer in order to build the main conceptualization and to fill its role.</Paragraph> <Paragraph position="3"> Also MARGIE(Schank,'75) makes some use of syntactic knowledge distributed among lexical definitions of words. This solution gives the entire control to the semantic interpreter and no syntactic functional representation is used. It seems,however, that an intermediate step, keeping the syntactic output separate from the semantic one, has the advantage of avoiding the multiplication of single pieces of syntactic knowledge. It also provides a simpler way of dealing with syntactic variants of the same sentence and a help in identifying coreferences.</Paragraph> <Paragraph position="4"> The semantic interpreter works fundamentally bottom-up and, although much is still to be attempted, it seems that it can usefully cooperate with a top-down parser to find the correct interpretation. These practical advantages will be taken into account also in the future development of the system. In fact it seems that, although no definite solution has been given to many linguistic problems, the interaction between two fully developped mechanisms controlling each other can provide an indication and a frame into which a more compact system can be built.</Paragraph> <Paragraph position="5"> In the present version of the system the interaction between the two modules is strictly sequential. In a more compact analyzer, syntactic specialists, i.e. simplified pieces of grammar specialized in particular syntactic phenomena, will be called by semantic interpreter according to opportunity. This second version is still being designed.</Paragraph> </Section> <Section position="11" start_page="181" end_page="181" type="metho"> <SectionTitle> VI ACKNOWLEDGEMENTS </SectionTitle> <Paragraph position="0"> The autors would like to thank Dr. Lina Massone for her contributions and assistance in the preparation of this paper.</Paragraph> </Section> class="xml-element"></Paper>