| { |
| "paper_id": "E85-1015", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T11:30:28.249349Z" |
| }, |
| "title": "LANGUAGE-BASED ENVIRONMENT FOR NATURAL LANGUAGE PARSING", |
| "authors": [ |
| { |
| "first": "A", |
| "middle": [], |
| "last": "Lehtola", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Helsinki University of Technology Helsinki", |
| "location": { |
| "country": "Finland" |
| } |
| }, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "", |
| "pdf_parse": { |
| "paper_id": "E85-1015", |
| "_pdf_hash": "", |
| "abstract": [], |
| "body_text": [ |
| { |
| "text": ").", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "The left The righ constituent constituent stack stack The syntax of these declarations can be seen in Figure 3 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 102, |
| "end": 111, |
| "text": "Figure 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "All properties of constituents may be referred in a uniform manner using their values straight.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "The system automatically takes into account the computational details associated to property types. For example, the system is automatically tuned to notice the inheritance of properties in their hierarchies.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "Extensive support to multidimensional analysis has been one of the central objectives in the design of the DPL-formalism.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "Patterning can be done in multiple dimensions and the property set associated to constituents can easily be extended.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "An example of a constituent structure and its property definitions is given in Figure 4 . In Figure 6 is The direction of a state (see Figure 2 .) selects the dependent candidate normally as L1 or R1.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 79, |
| "end": 87, |
| "text": "Figure 4", |
| "ref_id": "FIGREF3" |
| }, |
| { |
| "start": 93, |
| "end": 104, |
| "text": "Figure 6 is", |
| "ref_id": null |
| }, |
| { |
| "start": 135, |
| "end": 145, |
| "text": "Figure 2", |
| "ref_id": "FIGREF1" |
| } |
| ], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "A switch of state takes place by an assignment in the same way as linguistic properties are assigned.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "As an example the node V? of Figure 2 is defined formally in Figure 8 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 29, |
| "end": 38, |
| "text": "Figure 2", |
| "ref_id": "FIGREF1" |
| }, |
| { |
| "start": 62, |
| "end": 70, |
| "text": "Figure 8", |
| "ref_id": "FIGREF5" |
| } |
| ], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "More linguistically oriented argumentation of the DPL-formalism appears elsewhere (Nelimarkka, 1984a, and Nelimarkka, 1984b) .", |
| "cite_spans": [ |
| { |
| "start": 82, |
| "end": 105, |
| "text": "(Nelimarkka, 1984a, and", |
| "ref_id": null |
| }, |
| { |
| "start": 106, |
| "end": 124, |
| "text": "Nelimarkka, 1984b)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "R3", |
| "sec_num": null |
| }, |
| { |
| "text": "The architecture of the DPL-environment is described schematically in Figure 9 . The main parts are highlighted by heavy lines. The compiler produces highly optimized code (Lehtola, 1984) . Internally data structures are only partly dynamic for the reason of fast information fetch.", |
| "cite_spans": [ |
| { |
| "start": 172, |
| "end": 187, |
| "text": "(Lehtola, 1984)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 70, |
| "end": 78, |
| "text": "Figure 9", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "Ambiguities are expressed locally to minimize redundant search. The principle of structure sharing is followed whenever new data structures are built. In the manipulation of constituent structures there exists a special service routine for each combination of property and predication types. These routines take special care of time and memory consumption.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "For instance with regard replacements and insertions the copying includes physically only the path from the root of the list structure to the changed sublist.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "The logically shared parts will \u2022 be shared also physically. This stipulation minimizes memory usage.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "In the state transition network level the search is done depth first.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "To handle ambiquities DPL-functions and -relations process all alternative interpretations in parallel.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "In fact the alternatives are stored in the stacks and in the C-register as trees of alternants.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "In the first version of the DPL-compiler the generation rules were intermixed with the compiler code.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "The maintenance of the compiler grew harder when we experimented with new computational features. We parser facility lexicon maintenance information extraction system with graphic output Figure 9 . The architecture of the DPL-environment therefore started to develop a metacompiler in which compilation is defined by rules. At moment we are testing it and soon it will be in everyday use.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 187, |
| "end": 195, |
| "text": "Figure 9", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "The amount of LISP-code has greatly reduced with the rule based approach, and we are now planning to install the DPL-environment into IBM PC.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "Our parsers were aimed to be practical tools in real production applications. It was hence important to make the produced programs transferable.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| }, |
| { |
| "text": "As ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "THE ARCHITECTURE OF THE DPL-ENVIRONMENT", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": " Figure I0 . A trace of parsing process the parsing process described in Figure i0 . ", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 1, |
| "end": 10, |
| "text": "Figure I0", |
| "ref_id": null |
| }, |
| { |
| "start": 73, |
| "end": 82, |
| "text": "Figure i0", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "annex", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": {}, |
| "ref_entries": { |
| "FIGREF0": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "Figure I. The working storage of DPL-parsers" |
| }, |
| "FIGREF1": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "A two-way automaton for Finnish verbs" |
| }, |
| "FIGREF2": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "( PROPERTY: <type name> <possible values> ) : ( FEATURE: <type name> <possible values> ) ( CATEGORY: <type name> < <node definition>.. > ) ::= < <default value> <unique lisp atom>.. > ::= NoDefault : <unique lisp atom> ::= ( <node name> <feature set> <father node> ) ::= <unique lisp atom> ::= ( <feature value> ) : <empty> ::= / <name of an already defined node> : <empty> ::=Figure 3. The syntax of constituent structure and property definitions" |
| }, |
| "FIGREF3": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "An example of a constituent structure specification and the definition of an category tree implicit on the right and a double arrow (=>) terminates the execution at the first successful operation." |
| }, |
| "FIGREF4": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "The syntax of DPL-functions and DPL-relations A realisation of Subject <state in autom.>::= ( STATE: <state name> <direction> <state expr>.. ) <direction> ::= LEFT | RIGHT <state expr> ::= ( <lhs of s. expr> <impl> <state expr>.. ) ( <lhs of s. expr> <impl> <state change> ) <lhs of s. expr> ::= <function name> ~ <predicate expr>.-Phrase) -> (BUILD-PHRASE-ON RIGHT (C := V?)))" |
| }, |
| "FIGREF5": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "The expression of V? inFigure 2" |
| }, |
| "FIGREF6": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "." |
| }, |
| "FIGREF8": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "Default valuen -Phrase Associated values: (+Declarative -Declarative +Main -Main +Nominal -Nominal +Phrase -Phrase +Predicative -Predicative +Relative -Relative +Sentence -Sentence) Associated ~uncti onsl (C~nstFeat/INIT ConstFeat/FN CenstFeatl= ConstFeat/=:= ConstFeat/:-ConstFeat/,-/C CanstFeat/:= ConstFeat/:=/C) Figure ii. An example of information extraction utilities" |
| } |
| } |
| } |
| } |
