{ "paper_id": "M95-1024", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T03:12:42.584052Z" }, "title": "", "authors": [], "year": "", "venue": null, "identifiers": {}, "abstract": "", "pdf_parse": { "paper_id": "M95-1024", "_pdf_hash": "", "abstract": [], "body_text": [ { "text": "The Named Entity task consists of three subtasks (entity names, temporal expressions, number expressions) . The expressions to be annotated are \"unique identifiers\" of entities (organizations, persons, locations) , times (dates , times), and quantities (monetary values, percentages) .", "cite_spans": [ { "start": 177, "end": 212, "text": "(organizations, persons, locations)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "For many text processing systems, such identifiers are recognized primarily using local pattern-matchin g techniques . The TEI (Text Encoding Initiative) Guidelines for Electronic Text Encoding and Interchange cove r such identifiers (plus abbreviations) together in section 6 .4 and explain that the identifiers comprise \"textual feature s which it is often convenient to distinguish from their surrounding text . Names, dates and numbers are likely to be of particular importance to the scholar treating a text as source for a database ; distinguishing such items from the surrounding text is however equally important to the scholar primarily interested in lexis . \"", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "The task is to identify all instances of the three types of expressions in each text in the test set and to subcategorize the expressions . The original texts contain some SGML tags already ; the Named Entity task is to be performe d within the text delimited by the TXT, HL, DATELINE, and DD tags . (Note, however, that the DD tag sometimes doesn't appear at all, sometimes appears once, and sometimes appears twice . When it appears twice, only th e second instance is to be marked up for the Named Entity task . )", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "Lines within the portion of the article that start with the \"@\" sign signify a table or other special lin e formatting within the text and should NOT be annotated . (However, such lines may also appear within the portion of the article, and these should be annotated .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "The system must produce a single, unambiguous output for any relevant string in the text ; thus, this evaluation i s not based on a view of a pipelined system architecture in which Named Entity recognition would be completel y handled as a preprocess to sentence and discourse analysis . The task requires that the system recognize what a string represents, not just its superficial appearance . Sometimes, the right answer is superficially apparent, as in th e case of most, if not all, NUMEX expressions, and can be obtained by local pattern-matching techniques . In other cases, the right answer is not superficially apparent, as when a single capitalized word could represent the name of a location, person, or organization, and the answer may have to be obtained using techniques that draw informatio n from a larger context or from reference lists .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "The three subtasks correspond to three SGML tag elements : ENAMEX, TIMEX, and NUMEX. Th e subcategorization is captured by a SGML tag attribute called TYPE, which is defined to have a different set o f possible values for each tag element . The markup is described in section 2, below .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "Scoring of this task will be done using the same kinds of metrics that are used for scoring template-filling (information extraction) tasks . For specific information on the scoring, refer to \"MUC-6 Scoring System User' s Manual,\" prepared for MUC-6 by SAIC .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Performance Evaluation", "sec_num": "1" }, { "text": "Cumulative scores will be generated at several levels of description of the task, e .g . ,", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Performance Evaluation", "sec_num": "1" }, { "text": "\u2022 across subtasks ,", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Performance Evaluation", "sec_num": "1" }, { "text": "\u2022 for each subtask ,", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Performance Evaluation", "sec_num": "1" }, { "text": "\u2022 for the subcategorization aspect of each subtask ,", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Performance Evaluation", "sec_num": "1" }, { "text": "\u2022 for each part of the article that is included in the task (,,
, ) .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Performance Evaluation", "sec_num": "1" }, { "text": "The output of the systems to be evaluated will be in the form of SGML text markup . The only insertions allowed during tagging are tags enclosed in angled brackets . No extra whitespace or carriage returns are to be inserted ; otherwise, the offset count would change, which would adversely affect scoring .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Markup Description", "sec_num": "2" }, { "text": "The markup will have the following form :", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Markup Description", "sec_num": "2" }, { "text": "text-string Example :", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Markup Description", "sec_num": "2" }, { "text": "Taga Co .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Markup Description", "sec_num": "2" }, { "text": "The markup is defined in SGML Document Type Descriptions (DTDs), written for MUC-6 use by personnel at MITRE and maintained by personnel at NRaD . The DTDs enable annotators and system developers to use SGML validation tools to check the correctness of the SGML-tagged texts produced by the annotator or the system . The validation tools are available to MUC-6 participants in the file called muc6-sgml-tools . Annotators are using a software tool provided for MUC-6 by SRA Corporation to assist in generating the answer keys to be used fo r system training and testing .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Markup Description", "sec_num": "2" }, { "text": "This subtask is limited to proper names, acronyms, and perhaps miscellaneous other unique identifiers, which ar e categorized via the TYPE attribute as follows :", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Named Entities (ENAMEX tag element )", "sec_num": "2.2" }, { "text": "ORGANIZATION: named corporate, governmental, or other organizational entity PERSON : named person or famil y LOCATION : name of politically or geographically defined location (cities, provinces, countries, international regions, bodies of water, mountains, etc . )", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Named Entities (ENAMEX tag element )", "sec_num": "2.2" }, { "text": "This subtask is for \"absolute\" temporal expressions only ; explanation is provided in appendix B . The tagged tokens are categorized via the TYPE attribute as follows :", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Temporal Expressions (TIMEX tag element )", "sec_num": "2.3" }, { "text": "DATE : complete or partial date expressio n TIME : complete or partial expression of time of da y", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Temporal Expressions (TIMEX tag element )", "sec_num": "2.3" }, { "text": "This subtask is for two useful types of numeric expressions, monetary expressions and percentages . The numbers may be expressed in either numeric or alphabetic form .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Number Expressions (NUMEX tag element)", "sec_num": "2.4" }, { "text": "The task covers the complete expression, which is categorized via the TYPE attribute as follows :", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Number Expressions (NUMEX tag element)", "sec_num": "2.4" }, { "text": "MONEY : monetary expressio n PERCENT : percentag e", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Number Expressions (NUMEX tag element)", "sec_num": "2.4" }, { "text": "A vertical bar is being used to separate alternative TYPE attribute values in the answer key .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressing Alternative Attribute Value s", "sec_num": "3" }, { "text": "Alternative values will be given when the annotator does not have enough information to make a uniqu e categorization, even considering the context and the annotator's knowledge of the world .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressing Alternative Attribute Value s", "sec_num": "3" }, { "text": "When it is not certain that a string should be marked up, the annotator will include the STATUS attribute in th e markup to indicate that the markup is optional . The only value of the STATUS attribute is \"OPT .\" Examples of it s possible use can be found in the appendices, such as in appendix B (holiday names) .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Expressing Optional Markup (STATUS Attribute)", "sec_num": "3" }, { "text": "The ALT attribute will be used when the tagged string contains one or more substrings that should be considere d correct for the purposes of scoring the system response . Certain premodifiers (\"a,\" \"an,\" and \"the\") are automatically ignored by the scoring program (via the \"configuration\" file) and thus do not need to be speciall y marked in the key.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".Expressing Alternative String Boundaries (ALT Attribute )", "sec_num": "3" }, { "text": "The ALT attribute will be used sparingly . A possible TIMEX example is shown below .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".Expressing Alternative String Boundaries (ALT Attribute )", "sec_num": "3" }, { "text": "\"all of 1987 \" all of 1987", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".Expressing Alternative String Boundaries (ALT Attribute )", "sec_num": "3" }, { "text": "Multi-name expressions containing conjoined modifiers (with elision of the head of one conjunct) should be marke d up as separate expressions .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressions Involving Elision", "sec_num": "4" }, { "text": "\"North and South America \" North an d South America A similar case involving elision with number expressions :", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressions Involving Elision", "sec_num": "4" }, { "text": "\"10-and 20-dollar bills \" 10-an d 20-dollar bill s", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressions Involving Elision", "sec_num": "4" }, { "text": "In contrast, there is no elision in the case of single-name expressions containing conjoined modifiers ; such expressions should be marked up as a single expression .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressions Involving Elision", "sec_num": "4" }, { "text": "\"U .S . Fish and Wildlife Service \" U .S . Fish and Wildlife Service", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressions Involving Elision", "sec_num": "4" }, { "text": "The subparts of range expressions should be marked up as separate expressions .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressions Involving Elision", "sec_num": "4" }, { "text": "\"175 to 180 million Canadian dollars \" 175 to 180 million Canadia n dollars \"the 1986-87 academic year\" the 1986-8 7 academic year", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".1 Expressions Involving Elision", "sec_num": "4" }, { "text": "The systems must incorporate certain tokenization conventions . These conventions are contained in a separat e document titled \"Tokenization Rules . \"", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Effects of Tokenization Conventions", "sec_num": "4" }, { "text": "The tokenization conventions for MUC-6 have an impact on the boundaries of the strings to be tagged . For example, the conventions call for treating possessive forms, e .g ., \"California's,\" as multiple tokens, unless there is a name such as \"McDonald's [burger company]\" that is inherently possessive . See the separate documentation titled \"Tokenization Rules\" for further information and examples .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".2 Effects of Tokenization Conventions", "sec_num": "4" }, { "text": "There are cases where LOCATION (ENAMEX) expressions are to be tagged when they occur within TIMEX an d NUMEX expressions . See Appendices B and C for a description of cases where markup of nested expressions ma y occur . Entity names that appear within ENAMEX tags are *not* to be tagged .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Nested Expression s", "sec_num": "4.3" }, { "text": "Al Guidelines That Pertain to All Three TYPEs A .1 .1 Entity names used as modifiers in complex NPs that are not proper names are to be tagged when it is clea r to the annotator from context or the annotator's knowledge of the world that the name is that of an organization o r person .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "APPENDIX A . ENAMEX : SPECIFIC GUIDELINES", "sec_num": null }, { "text": "\"Bridgestone profits \" Bridgestone profit s \"the Clinton government \" the Clinton governmen t In some cases, multiword strings that are proper names will contain entity name substrings ; such strings are not decomposable ; therefore, the substrings are not to be tagged . (See A .1 .2 re special cases involving prenominal modifiers of person identifiers . ) \"Arthur Anderson Consulting \" Arthur Anderson Consulting [no markup for \"Arthur Anderson\" alone ] \"Boston Chicken Corp . \" Boston Chicken Corp . [no markup for \"Boston\" alone ] \"U .S . Fish and Wildlife Service \" U .S . Fish and Wildlife Service [no markup for \"U .S .\" alone] \"Northern California \" Northern California [no markup for \"California\" alone] \"West Texas \" West Texas [no markup for \"Texas\" alone] \"Ford Taurus \" Ford Tauru s [no markup, not even for \"Ford\" ] \"Dow Jones Industrial Average \" Dow Jones Industrial Averag e [no markup, not even for \"Dow Jones\" ] \"West Texas Intermediate crude \" West Texas Intermediate crud e [no markup, not even for \"West Texas\" --this example differs from those in A .1 .1, in that \"West Texa s Intermediate\" is a name and can therefore not be decomposed ] A .1 .4 Uncapitalized, common-noun designators such as \"division\" in the phrase \"Chrysler division\" are *not* considered part of an entity name .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "APPENDIX A . ENAMEX : SPECIFIC GUIDELINES", "sec_num": null }, { "text": "\"Chrysler division \" Chrysler divisio n \"the Kennedy family \" the Kennedy famil y \"the Southwest region \" the Southwest regio n A .1 .5 In a possessive construction, the possessor and possessed ENAMEX substrings should be tagge d separately. \"the mayor who built Candlestick Park for the Giants \" the mayor who built Candlestick Park for the Giant s< /ENAMEX > [a sports team] \"NASDAQ \" NASDAQ [a stock exchange ] \"European Community \" European Community \"Labor Party \" Labor Party \"Machinists union \" Machinists unio n A .2 .3 Miscellaneous types of proper names referring to facilities (e .g ., factories, hotels, universities, airports , hospitals, churches) will be tagged as ORGANIZATION unless (1) the name is clearly the name of the structure / place and not of an organization, or (2) the name is known to be the name of an organization but is used only i n reference to the facility as a structure/place . Some of the tagged strings may be marked optional in the answer ke y if the annotator thinks that it's highly unlikely (but not inconceivable) that the name represents an organization . ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "APPENDIX A . ENAMEX : SPECIFIC GUIDELINES", "sec_num": null }, { "text": "A .3 .1 Titles such as \"Mr.\" and role names such as \"President\" are *not* considered part of a person name. However, appositives such as \"Jr.\" *are* considered part of a person name . \"Mr . Harry Schearer\" Mr . Harry Schearer \"Secretary Robert Mosbacher \" Secretary Robert Mosbacher \"John Doe, Jr. \" John Doe, Jr . A .3 .2 Family names are to be tagged . \"the Kennedy family \" the Kennedy famil y \"the Kennedys \" the Kennedys [alternate form of identification of the Kennedy family entity A.3 .3 Miscellaneous types of proper names that are not to be tagged as PERSON include political groups, law s named after people, diseases/prizes named after people, saints .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "A.3 Guidelines That Pertain Only to PERSON", "sec_num": null }, { "text": "\"The Republicans held a rally . \" The Republicans held a rally . ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "A.3 Guidelines That Pertain Only to PERSON", "sec_num": null }, { "text": "A.4 .1 The phrase \"of \" following an organization name may or may not be part of the organization name proper. The annotation in the answer key will follow these guidelines : (1) If there is a corporate designator, i t marks the end of the organization name ; (2) if there is no corporate designator, the \"of \" is part of the organization name .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "A.4 Guidelines That Pertain Only to LOCATIO N", "sec_num": null }, { "text": "\"Hyundai of Korea, Inc . \" Hyundai of Korea, Inc . \"Hyundai, Inc . of Korea\" Hyundai, Inc . of Korea \"McDonald's of Japan \" McDonald's of Japan \"University of California in Los Angeles \" University of California in Los Angeles [where the first tagged string functions as organization rather than facility (see A .2, above, re facilities) --note also that the locative phrase \"in Los Angeles\" is not included inside the organization tag ] A .4 .2 Compound expressions in which place names are separated by a comma are to be tagged as separat e instances of LOCATION. \"Kaohsiung, Taiwan \" Kaohsiung, Taiwan \"Washington, D .C . \" Washington, D .C . A .4.3 Miscellaneous location-related strings that are not to be tagged as LOCATION include street addresses , street names, adjectival forms of location names . ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "A.4 Guidelines That Pertain Only to LOCATIO N", "sec_num": null } ], "back_matter": [ { "text": "Appendix C : Named Entity Task Definition (v2.1) APPENDIX B . TIMEX : SPECIFIC GUIDELINES B .1 Only \"absolute\" time expressions are to be tagged . To be considered an absolute time expression, th e expression must indicate a specific segment of time, as follows :\u2022 An expression of minutes must indicate a particular minute and hour, such as \"20 minutes after 10\" (not \"a fe w minutes after the hour,\" \"20 minutes after the hour\") .\u2022 An expression of hours must indicate a particular hour, such as \"midnight,\" \"twelve o'clock noon,\" \"noon\" (no t \"mid-day,\" \"morning\") .\u2022 An expression of days must indicate a particular day, such as \"Monday,\" \"10th of October\" (not \"first day o f the month\") .\u2022 An expression of seasons must indicate a particular season, such as \"autumn\" (not \"next season\").\u2022 An expression of financial quarters or halves of the year must indicate which quarter or half, such as \"fourt h quarter,\" \"first half.\" Note that there are no proper names, per se, representing these time periods . Nonetheless, these types of time expressions are important in the business domain and are therefore to be tagged.\u2022 An expression of years must indicate a particular year, such as \"1995\" (not \"the current year\") .\u2022 An expression of decades must indicate a particular decade, such as \"1980s\" (not \"the last 10 years\") .\u2022 An expression of centuries must indicate a particular century, such as \"the 20th century\" (not \"this century\") .B .2 Absolute time expressions are to be tagged as a single item. Subparts are not to be tagged separately, even in the case of possessive constructions .\"twelve o'clock noon \" twelve o'clock noon \"5 p .m . EST\" 5 p .m . EST \"January 1990 \" January 1990 \"fiscal 1989 \" fiscal 1989 \"the autumn report \" the autumn repor t \"an Indian summer of the soul \" an Indian summer of the sou l \"third quarter of 1991 \" third quarter of 1991 \"the fourth quarter ended Sept. 30 \" the fourth quarter ended Sept . 30 \"the three months ended Sept . 30\" [as paraphrase of \"the fourth quarter ended Sept . 30\" ] the three months ended Sept . 30 \"the first half of fiscal 1990 \" the first half of fiscal 1990 \"first-half profit \" first-half profi t \"fiscal 1989's fourth quarter \" fiscal 1989's fourth quarter \"4th period\" [of a year ] 4th period \"1975 World Series \" 1975 World Serie s B .3 Relative time expressions are not to be tagged, but any absolute times expressed as part of the entir e expression are to be tagged .\"July last year \" July last yea r \"this June \" this June \"next summer \" next summer \"thirty days before the end of the year \" thirty days before the end of the yea r [no markup ] \"the end of 1991 \" the end of 1991 [interpretation of \"end\" is unclear ; it could be a day or month, or something else ] B .4 Special days, such as holidays, that are referenced by name, are not required to be tagged ; they will b e marked optional in the answer key.\"because of the observance of All Saints' Day \" because of the observance of All Saints ' Day B .5 Rarely, multiword strings that are to be tagged as TIMEX will contain LOCATION (ENAMEX ) substrings that are also to be tagged.\"1 :30 p .m . Chicago time \" 1 :30 p .m . Chicago time APPENDIX C . NUMEX : SPECIFIC GUIDELINES C.1 The entire string expressing the monetary or percentage value is to be tagged.\"20 million New Pesos \" 20 million New Pesos \"$42 .1 million \" $42 .1 million \"million-dollar conferences \" million-dollar conference s \"15 pct \" 15 pct C .2 Juxtaposed strings expressing monetary values in two different currencies are to be tagged separately .\"#26 million ($43 .6 million) \" < NUMEX TYPE=\"MONEY\">#26 million ($43 . 6 million ) C .3 Modifying words that indicate the approximate value of a number or a \"relative position\" to a numbe r are to be excluded from the NUMEX tag . However, borderline cases surh as the quantifier \"several \" may be optionally included in the tagged string .\"about 5% \" about 5% \"over $90,000 \" over $90,000 \"several million dollars \" several million dollars C .4 Rarely, multiword strings that are to be tagged as NUMEX will contain LOCATION (ENAMEX ) substrings that are also to be tagged .\"the equivalent of less than a U .S . penny \" the equivalent of less than a U .S . penny \"20 million New Taiwan Dollars \" 20 million New Taiwan Dollars \"US$43 .6 million \" US$43 . 6 million C .5 Numeric expressions that do not use currency terms to indicate money values and that do not us e percentage terms to indicate percentages are not to be tagged . ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "annex", "sec_num": null } ], "bib_entries": {}, "ref_entries": { "TABREF0": { "text": "", "num": null, "html": null, "content": "
1 INTRODUCTION
1 .1 Scop e
(Version 2 .1, 27 JUL 95 )
1INTRODUCTION
1 .1Scope
1 .2Performance Evaluation
2 TASK OVERVIEW
2 .1Markup Description
2 .2Named Entities (ENAMEX tag element)
2 .3Temporal Expressions (TIMEX tag element)
2 .4Number Expressions (NUMEX tag element)
3 NOTATION RESERVED FOR USE IN THE ANSWER KEYS
3 .1Expressing Alternative Attribute Values
3 .2Expressing Optional Markup (STATUS Attribute)
3 .3Expressing Alternative String Boundaries (ALT Attribute)
4 GUIDELINES FOR MARKUP OF EXCEPTIONAL CONSTRUCTIONS
4 .1Expressions Involving Elision
4 .2Effects of Tokenization Conventions
4 .3Nested Expressions
APPENDIX A . ENAMEX : SPECIFIC GUIDELINES
APPENDIX B . TIMEX : SPECIFIC GUIDELINES1
APPENDIX C . NUMEX : SPECIFIC GUIDELINES1
", "type_str": "table" }, "TABREF3": { "text": "Finger Lakes Area Hospital Corp . \" Finger Lakes Area Hospital Corp .", "num": null, "html": null, "content": "
\"Four Seasons Hotels \"
[markup as ORGANIZATION likely ]
\"Unification Church\"
[markup as ORGANIZATION likely ]
\"the White House \"
[markup as ORGANIZATION, or no markup]
\"Trinity Lutheran Church\"
[markup as ORGANIZATION, or no markup]
\"General Hospital \"
[markup as ORGANIZATION, or no markup]
", "type_str": "table" } } } }