| { | |
| "paper_id": "W14-0145", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T05:51:30.238129Z" | |
| }, | |
| "title": "Semi-Automatic Extension of Sanskrit Wordnet using Bilingual Dictionary", | |
| "authors": [ | |
| { | |
| "first": "Sudha", | |
| "middle": [], | |
| "last": "Bhingardive", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Indian Institute of Technology", | |
| "location": { | |
| "settlement": "Bombay" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Tanuja", | |
| "middle": [], | |
| "last": "Ajotikar", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Indian Institute of Technology", | |
| "location": { | |
| "settlement": "Bombay" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Irawati", | |
| "middle": [], | |
| "last": "Kulkarni", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Indian Institute of Technology", | |
| "location": { | |
| "settlement": "Bombay" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Malhar", | |
| "middle": [], | |
| "last": "Kulkarni", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Indian Institute Technology", | |
| "location": { | |
| "settlement": "Bombay" | |
| } | |
| }, | |
| "email": "[email protected]" | |
| }, | |
| { | |
| "first": "Pushpak", | |
| "middle": [], | |
| "last": "Bhattacharyya", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "In this paper, we report our methods and results of using, for the first time, semi-automatic approach to enhance an Indian language Wordnet. We apply our methods to enhancing an already existing Sanskrit Wordnet created from Hindi Wordnet (which is created from Princeton Wordnet) using expansion approach. We base our experiment on an existing bilingual Sanskrit English Dictionary and show how lemma in this dictionary can be mapped to Princeton Wordnet through which corresponding Sanskrit synsets can be populated by Sanskrit lexemes. This our method will also show how absence of resources of a pair of languages need not be an obstacle, if another resource of one of them is available. Sanskrit being historically related to languages of Indo-European family, we believe that this semi-automatic approach will help enhance Wordnets of other Indian languages of the same family.", | |
| "pdf_parse": { | |
| "paper_id": "W14-0145", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "In this paper, we report our methods and results of using, for the first time, semi-automatic approach to enhance an Indian language Wordnet. We apply our methods to enhancing an already existing Sanskrit Wordnet created from Hindi Wordnet (which is created from Princeton Wordnet) using expansion approach. We base our experiment on an existing bilingual Sanskrit English Dictionary and show how lemma in this dictionary can be mapped to Princeton Wordnet through which corresponding Sanskrit synsets can be populated by Sanskrit lexemes. This our method will also show how absence of resources of a pair of languages need not be an obstacle, if another resource of one of them is available. Sanskrit being historically related to languages of Indo-European family, we believe that this semi-automatic approach will help enhance Wordnets of other Indian languages of the same family.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "Wordnet is a lexical semantic network, widely used in various applications of natural language processing. Princeton wordnet (PWN) is the mother of all Wordnets (Fellbaum, 1988) . It was created at the Cognitive Science Laboratory of Princeton University. EuroWordNet (Vossen, 1998; Vossen, 2000) , CoreNet (Choi, 2004) , In-doWordNet (Bhattacharyya, 2010) , HowNet (Zhendong, 2000) , MultiWordNet , BabelNet (Navigli, 2012) and so many other Nets are also some of the most commonly used semantic networks.", | |
| "cite_spans": [ | |
| { | |
| "start": 161, | |
| "end": 177, | |
| "text": "(Fellbaum, 1988)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 268, | |
| "end": 282, | |
| "text": "(Vossen, 1998;", | |
| "ref_id": "BIBREF15" | |
| }, | |
| { | |
| "start": 283, | |
| "end": 296, | |
| "text": "Vossen, 2000)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 299, | |
| "end": 319, | |
| "text": "CoreNet (Choi, 2004)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 335, | |
| "end": 356, | |
| "text": "(Bhattacharyya, 2010)", | |
| "ref_id": "BIBREF2" | |
| }, | |
| { | |
| "start": 366, | |
| "end": 382, | |
| "text": "(Zhendong, 2000)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 409, | |
| "end": 424, | |
| "text": "(Navigli, 2012)", | |
| "ref_id": "BIBREF10" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "PWN is manually created using the knowledge from various dictionaries. Several Wordnets are created semi-automatically using the expansion approach from PWN. Many of them use bilingual dictionaries or Wikipedia. This type of creation saves enormous manual efforts and time. However, it demands high quality machinereadable resources in the respective languages.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Sanskrit wordnet (SWN) (Kulkarni et.al, 2010) is manually created using the expansion approach from Hindi wordnet (HWN), which in turn was created from the Princeton Wordnet. The current status of Sanskrit wordnet is stated in ", | |
| "cite_spans": [ | |
| { | |
| "start": 23, | |
| "end": 45, | |
| "text": "(Kulkarni et.al, 2010)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "In this work, we aim to report our experiences to populate SWN by a semi-automated approach. Currently, manual approach is used which is time consuming and tedious. Following are the reasons that make manual approach time consuming.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "In the available lexical literature of Sanskrit (given below in Section 3), normal range for number of words in any synset varies between 1-20 e.g., l\u012bl\u0101 [a game (6 synset members)], v\u1e5bddha\u1e25 [an old man (20 synset members)], bhak\u1e63a\u1e47am [an act of eating (20 synset members)]. Synsets with only one word are common in the cases of coined words, instrument names and kinship relations. However, some synsets exceed this limit and have huge number of words as its members. We note below some of the prominent phenomena. For creating above mentioned synsets, lexicographers gathered information from various resources, e.g., while creating a concept of yuddha (a war), 97 words were collected from various lex-ical resources given below: Spoken Sanskrit Dictionary 1 (7 words), Apate's Sanskrit-English Dictionar 2 (7 words), Monier William's English-Sanskrit Dictionary 3 (57 words) and Shabdakalpadrum (80 words).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Large number of synonyms for a Sanskrit word", | |
| "sec_num": "2.1" | |
| }, | |
| { | |
| "text": "After collecting the words, duplicate words were eliminated. Words representing proper meanings are entered in the synset. This process is monetarily expensive and time consuming. Automatic approach can help populate such synsets using bilingual dictionaries. In the process there will be over-generation which will have to be controlled by manual approach.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Large number of synonyms for a Sanskrit word", | |
| "sec_num": "2.1" | |
| }, | |
| { | |
| "text": "While creating the synsets, appropriate selection of words is required to express the precise meaning. In Hindu texts, which are mainly in Sanskrit there are various names for a single deity e.g., Vi\u1e63\u1e47u (Hindu deity) has 132 names, K\u1e5b\u1e63\u1e47a has 132 names and R\u0101ma has 67 names. For creating synsets of these deities one must be very careful as K\u1e5b\u1e63\u1e47a and R\u0101ma are incarnations of Vi\u1e63\u1e47u and can easily get interchanged and thereby affecting the intended meaning.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Appropriate selection of words for creating synsets", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "The road-map of the paper is as follows. Section 3 presents the related work. Section 4 explains the methodology used for extension of SWN. Section 5 illustrates results. Outcomes are presented in Section 6. Section 7 includes conclusion and future work.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Appropriate selection of words for creating synsets", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "Most of the Wordnets are created by expansion approach using PWN. Several Wordnets have tried to increase their coverage using various automatic or semi-automatic approaches. Some of them are listed below. CoreNet (Choi, 2004) is an automatically constructed Wordnet, which uses a Japanese-Korean electronic dictionary. Korean words are programmatically generated during translation from Japanese. BabelNet (Navigli, 2012 ) is a very large, wide-coverage, multilingual semantic network. This resource is created by mapping a multilingual encyclopedic knowledge repository (Wikipedia) and a computational lexicon of English (PWN). The integration is performed via an automatic mapping and by filling in lexical gaps in resource-poor languages with the aid of Machine Translation. This provides concepts and named entities, lexicalized in many languages and connected with large amounts of semantic relations. Chinese Wordnet (Renjie Xu, 2008) is developed in an automatic manner by translating English words to Chinese using Chinese-English dictionary. Czech wordnet (Karel Pala, 2008) is automatically extended from PWN using machine-readable bilingual dictionary. Polish WordNet (M. Derwojedowa, 2008) is designed semi-automatically by extracting lexical relations from the large Polish corpora. Lexicographers are used for mapping these relations with PWN.", | |
| "cite_spans": [ | |
| { | |
| "start": 214, | |
| "end": 226, | |
| "text": "(Choi, 2004)", | |
| "ref_id": "BIBREF3" | |
| }, | |
| { | |
| "start": 407, | |
| "end": 421, | |
| "text": "(Navigli, 2012", | |
| "ref_id": "BIBREF10" | |
| }, | |
| { | |
| "start": 932, | |
| "end": 941, | |
| "text": "Xu, 2008)", | |
| "ref_id": "BIBREF16" | |
| }, | |
| { | |
| "start": 1066, | |
| "end": 1084, | |
| "text": "(Karel Pala, 2008)", | |
| "ref_id": "BIBREF11" | |
| }, | |
| { | |
| "start": 1184, | |
| "end": 1202, | |
| "text": "Derwojedowa, 2008)", | |
| "ref_id": "BIBREF4" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Related Work", | |
| "sec_num": "3" | |
| }, | |
| { | |
| "text": "English dictionary used for extending SWN?", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why was Monier William's Sanskrit-", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "We have used the publicly available Monier William's Sanskrit-English dictionary for SWN semiautomatic extension. The list of all the texts used by Monier Williams is publicly available. This dictionary includes over 1, 80, 000 words and definitions. All entries are organized according to the root of a word, the dhatu, which offers better understanding of the meaning of the word. It includes special references to cognate Indo-European languages as well as literary citations. It provides precise meanings for the words in the Vedic literature, which is useful for studying the scriptures. This is one of the most comprehensive and useful Sanskrit-English dictionaries. The other reason for using this dictionary for the present purpose, fortunately, is availability. Out of all the lexical resources mentioned above, only this is available in program readable format which makes this resource singularly important from the point of view of present research. One of the outputs of the use of this resource is extraction of proper nouns. We have automatically extracted them and added to SWN without linking them to PWN. This method is explained in Section 4.2.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why was Monier William's Sanskrit-", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "SWN is created by expansion approach from HWN, which was in turn created by PWN. Our selected resource is in Sanskrit and English.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Methodologies used for extending SWN", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "Therefore, in order to utilize it for the present purpose we have to link PWN directly to SWN. We link Sanskrit-English dictionary to PWN by using a heuristic. This will be automatic approach. These linkages are validated by lexicographers. This will be manual approach. Thus we will populate SWN by semi-automatic approach using this resource.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Methodologies used for extending SWN", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "William's dictionary contains Sanskrit words along with its English description. The description is concise for most of the Sanskrit words, e.g., kamala (lotus) has the description 'a lotus'. In comparison, PWN glosses are descriptive as shown in Figure 3 . Finding the maximum overlap between the description words in dictionary and PWN gloss words is not efficient as we get several possible mappings. It is monetarily expensive and time consuming to generate and validate these mappings. Therefore, this type of heuristic is not suitable for linking dictionary to PWN.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 247, | |
| "end": 255, | |
| "text": "Figure 3", | |
| "ref_id": "FIGREF2" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Heuristics used for linking William's dictionary to PWN", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "William's dictionary is a very rich resource in Sanskrit language, which is useful for extending the SWN. Hence, we linked dictionary to PWN using a heuristic, which finds the maximum overlap between description words in dictionary and words in PWN synsets. Using this heuristic, the dictionary entries are linked to PWN. We got 14653 single and 55059 multiple possible mappings. Lexicographers are in the process of validating these mappings. The architecture diagram of the process is shown in the steps for the procedure to link dictionary to PWN. \uf0b7 For a Sanskrit word Sw, from dictionary, its equivalent English description is taken and its maximum overlap with words in the PWN synsets is found.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Heuristics used for linking William's dictionary to PWN", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "\uf0b7 Sw is directly mapped to the synset if the word in the description is found to be monosemous in PWN.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Heuristics used for linking William's dictionary to PWN", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "\uf0b7 The mapping is evaluated manually if the word in the description is found to be polysemous in PWN.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Heuristics used for linking William's dictionary to PWN", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "After successful mapping, all Sanskrit words are added in SWN. This task can be explained with the help of an example, for the word 'kartr' (spinner), we found three possible mappings in PWN. For validating these multiple possible mappings, we designed an interface as shown in Figure 5 . It provides various functionalities on mappings viz., display, search, validate and delete. A lexicographer will select an appropriate mapping with the synset in PWN of correct sense.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 278, | |
| "end": 286, | |
| "text": "Figure 5", | |
| "ref_id": "FIGREF5" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Heuristics used for linking William's dictionary to PWN", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "After manual validation, all dictionary entries with valid mappings are inserted into the Sanskrit wordnet. Adding of all the dictionary entries maually requires excessive efforts. Thus, a semi-automatic approach will save these excessive manual efforts.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Heuristics used for linking William's dictionary to PWN", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "If the English description of the Sanskrit word began with the phrase 'Name of a', all such words can be considered as proper nouns. For example, the word 'Brahamhapuri' has the description, 'Name of a location'. Currently all proper nouns are part of the Wordnet. However, it is yet to be decided whether these are maintained in a separate gazetteer (gazetteers are those which contain entities themselves that are proper nouns), which will in turn link to SWN. If it is decided that they are to be treated as a part of Wordnet then it would add 14,339 synsets to SWN. Some of the extracted nouns are class names. For example, the word 'Ustika' has the description 'Name of a kind of plant' and the word 'Bhaumadevalipi' has the description 'Name of a kind writing'. Both these words are class names. All class names are not stored in a gazetteer. They are very much stored in the SWN. So far, fiftyfive class names are extracted from the dictionary and stored in SWN.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Other automatic application of William's dictionary to populate SWN", | |
| "sec_num": "4.2" | |
| }, | |
| { | |
| "text": "As discussed in Section 4.1 we are linking dictionary with PWN. There are 14, 653 Sanskrit words for which single mappings were found in PWN and 55, 059 words for which multiple mappings were found in PWN. The work of these mappings is still under validation process. We have extracted 14,339 proper nouns from dictionary, which are not covered by SWN.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Results", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "These proper nouns must get inserted into SWN as these are most frequent occurrences in Sanskrit literature. Current synset coverage status of SWN is illustrated in Table 1 . After adding dictionary entries, SWN coverage will increase considerably. With this semi-automatic approach, SWN will be a richer lexical resource in Sanskrit language. As this dictionary is considered as an authentic lexical resource for Sanskrit we can remove the word 'd\u012bpti\u1e25' from the corresponding SWN synset.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 165, | |
| "end": 172, | |
| "text": "Table 1", | |
| "ref_id": "TABREF0" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Results", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "6.2 Coverage of HWN will also improve with the help of dictionary. For example, dictionary provides the same English meaning 'moonless' for all the Sanskrit words namely 'acandra', 'na\u1e63\u1e6dacandra', 'ni\u1e25somaka', and 'visoma'. In the existing HWN, the concept of 'moonless' is not available. It is also not covered in SWN as it is created using expansion approach from HWN. The above mentioned words form a synset and can be added to SWN and then be further borrowed in HWN. In this way, we are also increasing the HWN coverage using dictionary and SWN as shown below. ", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Results", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "We have attempted to implement a semi-automatic approach for Sanskrit wordnet extension using Monier William's Sanskrit-English dictionary. Dictionary entries are automatically extracted and linked to PWN which need manual validation. For this purpose we have created a tool ( Figure 5 ) which is language independent and therefore can be adopted by other similar language pairs. Post manual validation, all these entries will be inserted to SWN. Also, we have automatically extracted proper nouns from dictionary, which play an important role in Sanskrit literature. With the help of this approach we are correcting existing synset members of SWN and existing SWN-HWN-PWN linkages. HWN coverage can also be increased with the help of this approach. Following this approach, we will generate all semantic and lexical relations automatically from the same bilingual dictionary. This work can be extended using other resources like B\u04e7htlingk and Roth's Sanskrit-German dictionary along with Monier William's dictionary for learning some useful patterns to make SWN a rich resource in Sanskrit language. ", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 277, | |
| "end": 285, | |
| "text": "Figure 5", | |
| "ref_id": "FIGREF5" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Conclusion and Future work", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "http://spokensanskrit.de/ 2 http://www.aa.tufs.ac.jp/~tjun/sktdic/", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "http://www.sanskrit-lexicon.uni-koeln.de/monier/mwauthorities/mwauth_SktDevaUnicode.html#rec-ord_Lalit_", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| } | |
| ], | |
| "back_matter": [ | |
| { | |
| "text": "We thank to Dr. Peter Scharf, President, Sanskrit Library, United States of America, for providing the database. We acknowledge the work which provides an XML encoding of the Monier Williams Sanskrit-English Dictionary which was done in collaboration between the Institute of Indology and Tamil Studies (IITS) of the University of Cologne and Brown University's Sanskrit Department.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Acknowledgements", | |
| "sec_num": null | |
| } | |
| ], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "Coping with lexical gaps when building aligned multilingual wordnets", | |
| "authors": [ | |
| { | |
| "first": "Luisa", | |
| "middle": [], | |
| "last": "Bentivogli", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Emanuele", | |
| "middle": [], | |
| "last": "Pianta", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Fabio", | |
| "middle": [], | |
| "last": "Pianesi", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2000, | |
| "venue": "Proceedings of LREC2000", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Luisa Bentivogli, Emanuele Pianta and Fabio Pianesi, 2000. Coping with lexical gaps when building aligned multilingual wordnets, In Proceedings of LREC2000, Athens, Greece.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Looking for lexical gaps", | |
| "authors": [ | |
| { | |
| "first": "Luisa", | |
| "middle": [], | |
| "last": "Bentivogli", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Emanuele", | |
| "middle": [], | |
| "last": "Pianta", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2000, | |
| "venue": "Proceedings of Euralex", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Luisa Bentivogli, Emanuele Pianta, 2000. Looking for lexical gaps, Proceedings of Euralex, Stuttgart, Ger- many.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Lexical Resources Engineering Conference (LREC)", | |
| "authors": [ | |
| { | |
| "first": "Pushpak", | |
| "middle": [], | |
| "last": "Bhattacharyya", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2010, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Pushpak Bhattacharyya, 2010. IndoWordNet, Lexical Resources Engineering Conference (LREC), Malta.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Procedures and Problems in Korean-Chinese-Japanese Wordnet with Shared Semantic Hierarchy", | |
| "authors": [ | |
| { | |
| "first": "Hee-Sook", | |
| "middle": [], | |
| "last": "Key-Sun Choi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Bae", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2004, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Key-Sun Choi and Hee-Sook Bae, 2004. Procedures and Problems in Korean-Chinese-Japanese Wordnet with Shared Semantic Hierarchy, GWC2004.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Words, concepts and relations in the construction of Polish WordNet", | |
| "authors": [ | |
| { | |
| "first": "Magdalena", | |
| "middle": [], | |
| "last": "Derwojedowa", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Maciej", | |
| "middle": [], | |
| "last": "Piasecki", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Stanisaw", | |
| "middle": [], | |
| "last": "Szpakowicz", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Magdalena", | |
| "middle": [], | |
| "last": "Zawisawska", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Bartosz", | |
| "middle": [], | |
| "last": "Broda", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "Proceedings of the Global WordNet Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Magdalena Derwojedowa, Maciej Piasecki, Stanisaw Szpakowicz, Magdalena Zawisawska and Bartosz Broda, 2008. Words, concepts and relations in the construction of Polish WordNet, In Proceedings of the Global WordNet Conference, Seged, Hungary.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Knowledge Description: What, How and Who?", | |
| "authors": [ | |
| { | |
| "first": "Dong", | |
| "middle": [], | |
| "last": "Dong Zhen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1988, | |
| "venue": "Proceedings of the International Symposium on Electronic Dictionaries, Tokyo", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Dong Zhen Dong, 1988. Knowledge Description: What, How and Who? , In Proceedings of the Inter- national Symposium on Electronic Dictionaries, To- kyo, Japan.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "WordNet: An Electronic Database", | |
| "authors": [ | |
| { | |
| "first": "Christiane", | |
| "middle": [], | |
| "last": "Fellbaum", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1998, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Christiane Fellbaum, 1998. WordNet: An Electronic Database, MIT Press, Cambridge, MA.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "Automatic Acquisition of Hyponyms from Large Text Corpora", | |
| "authors": [ | |
| { | |
| "first": "Marti", | |
| "middle": [], | |
| "last": "Hearst", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Proc. of International Conference on Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Marti Hearst, 1992. Automatic Acquisition of Hypo- nyms from Large Text Corpora, Proc. of Interna- tional Conference on Computational Linguistics, COLING1992.", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "Irawati Kulkarni", | |
| "authors": [ | |
| { | |
| "first": "Malhar", | |
| "middle": [], | |
| "last": "Kulkarni", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Chaitali", | |
| "middle": [], | |
| "last": "Dangarikar", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": null, | |
| "venue": "Abhishek Nanda and Pushpak Bhattacharyya, 2010. Introducing Sanskrit Wordnet, 5th International Conference on Global Wordnet (GWC2010)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Malhar Kulkarni, Chaitali Dangarikar, Irawati Kul- karni, Abhishek Nanda and Pushpak Bhattacharyya, 2010. Introducing Sanskrit Wordnet, 5th Interna- tional Conference on Global Wordnet (GWC2010), Mumbai.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "Lexicographic traditions in India and Sanskrit", | |
| "authors": [ | |
| { | |
| "first": "Malhar", | |
| "middle": [], | |
| "last": "Kulkarni", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "Journal of Language Technology", | |
| "volume": "", | |
| "issue": "1", | |
| "pages": "160--165", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Malhar Kulkarni, 2008. Lexicographic traditions in In- dia and Sanskrit, Journal of Language Technology, (1) pp. 160-165.", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "BabelNet: The Automatic Construction, Evaluation and Application of a Wide-Coverage Multilingual Semantic Network", | |
| "authors": [ | |
| { | |
| "first": "Roberto", | |
| "middle": [], | |
| "last": "Navigli", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Simone", | |
| "middle": [], | |
| "last": "Ponzetto", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2012, | |
| "venue": "Artificial Intelligence", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1016/j.artint.2012.07.001" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Roberto Navigli and Simone Ponzetto, 2012. BabelNet: The Automatic Construction, Evaluation and Application of a Wide-Coverage Multilingual Semantic Network, Artificial Intelligence, Elsevier.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "Semiautomatic Linking of New Czech Synsets Using Princeton WordNet, Intelligent Information Systems", | |
| "authors": [ | |
| { | |
| "first": "Karel", | |
| "middle": [], | |
| "last": "Pala", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Dana", | |
| "middle": [], | |
| "last": "Hlav\u00e1ckov\u00e1", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Va\u0161ek", | |
| "middle": [], | |
| "last": "", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Karel Pala, Dana Hlav\u00e1ckov\u00e1, and Va\u0161ek, 2008. Semi- automatic Linking of New Czech Synsets Using Princeton WordNet, Intelligent Information Sys- tems.", | |
| "links": null | |
| }, | |
| "BIBREF12": { | |
| "ref_id": "b12", | |
| "title": "History of Sanskrit Lexicography", | |
| "authors": [ | |
| { | |
| "first": "Patkar", | |
| "middle": [], | |
| "last": "Madhukar Mangesh", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1981, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Madhukar Mangesh Patkar, 1981. History of Sanskrit Lexicography, Munshiram Manoharlal Publishers, Delhi.", | |
| "links": null | |
| }, | |
| "BIBREF13": { | |
| "ref_id": "b13", | |
| "title": "Learning Dictionaries for Information Extraction using Multilevel Bootstrapping", | |
| "authors": [ | |
| { | |
| "first": "Ellen", | |
| "middle": [], | |
| "last": "Riloff", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Rosie", | |
| "middle": [], | |
| "last": "Jones", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1999, | |
| "venue": "Proc. of National Conference on Artificial Intelligence", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Ellen Riloff and Rosie Jones, 1999. Learning Diction- aries for Information Extraction using Multilevel Bootstrapping, Proc. of National Conference on Ar- tificial Intelligence.", | |
| "links": null | |
| }, | |
| "BIBREF14": { | |
| "ref_id": "b14", | |
| "title": "Euro WordNet General Document", | |
| "authors": [ | |
| { | |
| "first": "Piek", | |
| "middle": [], | |
| "last": "Vossen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2002, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Piek Vossen, 2002. Euro WordNet General Document, University of Amsterdam.", | |
| "links": null | |
| }, | |
| "BIBREF15": { | |
| "ref_id": "b15", | |
| "title": "EuroWordNet: A Multilingual Database with Lexical Semantic Networks", | |
| "authors": [ | |
| { | |
| "first": "Piek", | |
| "middle": [], | |
| "last": "Vossen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1998, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Piek Vossen, 1998. EuroWordNet: A Multilingual Da- tabase with Lexical Semantic Networks, Kluwer, Dordrecht, Netherlands.", | |
| "links": null | |
| }, | |
| "BIBREF16": { | |
| "ref_id": "b16", | |
| "title": "An Integrated Approach for Automatic Construction of Bilingual Chinese-English WordNet", | |
| "authors": [ | |
| { | |
| "first": "Renjie", | |
| "middle": [], | |
| "last": "Xu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Zhiqiang", | |
| "middle": [], | |
| "last": "Gao", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yingji", | |
| "middle": [], | |
| "last": "Pan", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yuzhong", | |
| "middle": [], | |
| "last": "Qu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Zhisheng", | |
| "middle": [], | |
| "last": "Huang", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "ASWC 2008", | |
| "volume": "5367", | |
| "issue": "", | |
| "pages": "302--314", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Renjie Xu, Zhiqiang Gao, Yingji Pan, Yuzhong Qu, Zhisheng Huang, 2008. An Integrated Approach for Automatic Construction of Bilingual Chinese-Eng- lish WordNet, ASWC 2008, LNCS 5367, 302-314.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "type_str": "figure", | |
| "num": null, | |
| "text": "SWN manual creation", | |
| "uris": null | |
| }, | |
| "FIGREF1": { | |
| "type_str": "figure", | |
| "num": null, | |
| "text": "SWN semi-automatic creation", | |
| "uris": null | |
| }, | |
| "FIGREF2": { | |
| "type_str": "figure", | |
| "num": null, | |
| "text": "Dictionary and PWN entry for kamala (a lotus)", | |
| "uris": null | |
| }, | |
| "FIGREF3": { | |
| "type_str": "figure", | |
| "num": null, | |
| "text": "", | |
| "uris": null | |
| }, | |
| "FIGREF4": { | |
| "type_str": "figure", | |
| "num": null, | |
| "text": "Architecture diagram", | |
| "uris": null | |
| }, | |
| "FIGREF5": { | |
| "type_str": "figure", | |
| "num": null, | |
| "text": "Interface for validating multiple possible mappings", | |
| "uris": null | |
| }, | |
| "FIGREF6": { | |
| "type_str": "figure", | |
| "num": null, | |
| "text": "HWN enriched with SWN and William's dictionary", | |
| "uris": null | |
| }, | |
| "TABREF0": { | |
| "html": null, | |
| "content": "<table><tr><td colspan=\"2\">POS Noun Verb</td><td>Ad-</td><td>Adjec-</td></tr><tr><td/><td/><td>verb</td><td>tive</td></tr><tr><td>synset</td><td>17413 1246</td><td>263</td><td>3990</td></tr><tr><td>counts</td><td/><td/><td/></tr></table>", | |
| "type_str": "table", | |
| "num": null, | |
| "text": "" | |
| }, | |
| "TABREF1": { | |
| "html": null, | |
| "content": "<table/>", | |
| "type_str": "table", | |
| "num": null, | |
| "text": "" | |
| }, | |
| "TABREF4": { | |
| "html": null, | |
| "content": "<table><tr><td>6.3 Some existing SWN synsets are not linked</td></tr><tr><td>with PWN as SWN-PWN linking is via</td></tr><tr><td>HWN. We are also improving these link-</td></tr><tr><td>ages using the dictionary. For example, an</td></tr><tr><td>HWN synset corresponding to one of the</td></tr><tr><td>synsets of vil\u0101sin in SWN, is not linked</td></tr><tr><td>with PWN. English description of vil\u0101sin is</td></tr><tr><td>given as 'coquettish' in the William's dic-</td></tr><tr><td>tionary. Both Sanskrit and English interpre-</td></tr><tr><td>tation are under the same POS category of</td></tr><tr><td>adjective.</td></tr></table>", | |
| "type_str": "table", | |
| "num": null, | |
| "text": "Thus, now we can link this SWN synset to PWN synset. In this way we are improving SWN-HWN-PWN linkages." | |
| } | |
| } | |
| } | |
| } |