| { |
| "paper_id": "2021", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T16:20:27.904589Z" |
| }, |
| "title": "Is this Enough? An evaluation of the Malayalam Wordnet", |
| "authors": [ |
| { |
| "first": "Nandu", |
| "middle": [ |
| "Chandran" |
| ], |
| "last": "Nair", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "University of Trento", |
| "location": {} |
| }, |
| "email": "" |
| }, |
| { |
| "first": "Maria-Chiara", |
| "middle": [], |
| "last": "Giangregorio", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "University of Trento", |
| "location": {} |
| }, |
| "email": "mariegiangregorio@protonmail.com" |
| }, |
| { |
| "first": "Fausto", |
| "middle": [], |
| "last": "Giunchiglia", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "University of Trento", |
| "location": {} |
| }, |
| "email": "fausto.giunchiglia@unitn.it" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "The quality of a product is the degree to which a product meets the Customer's expectations, which must also be valid in the case of lexical-semantic resources. Conducting a periodic evaluation of resources is essential to ensure that they meet a native speaker's expectations and are free from errors. This paper defines the possible errors that a lexical-semantic resource can contain, how they may impact downstream applications and explains the steps applied to evaluate and quantify the quality of Malayalam WordNet. Malayalam is one of the classical languages of India. We propose an approach allowing to subset the part of the WordNet tied to the lowest quality scores. We aim to work on this subset in a crowdsourcing context to improve the quality of the resource.", |
| "pdf_parse": { |
| "paper_id": "2021", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "The quality of a product is the degree to which a product meets the Customer's expectations, which must also be valid in the case of lexical-semantic resources. Conducting a periodic evaluation of resources is essential to ensure that they meet a native speaker's expectations and are free from errors. This paper defines the possible errors that a lexical-semantic resource can contain, how they may impact downstream applications and explains the steps applied to evaluate and quantify the quality of Malayalam WordNet. Malayalam is one of the classical languages of India. We propose an approach allowing to subset the part of the WordNet tied to the lowest quality scores. We aim to work on this subset in a crowdsourcing context to improve the quality of the resource.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "Internet is composed of machine-readable lexical or lexical semantic resources (Herring, 2008) . However, only a small percentage of languages have these resources developed in full scale (Herring, 2008) . Hence the majority of languages are classified as under-resourced (Besacier et al., 2014) . In our study, we consider a language to be under-resourced if it has the following characteristics: lack of computerreadable resources, lack of linguist experts, and limited usage on the internet (Krauwer, 2003) . The different large scale availability of language resources results in a digital language divide (Warschauer, 2002) . Figure 1 shows the visual representation of the study's findings on language-wise Internet users in 2021 and was conducted by KPMG India. We notice that usage of Hindi is greater than Malayalam. Bringing more languages online may ultimately be an exercise in cultural preservation, rather than utility. The probability of retrieving a relevant result using an under-resourced language is comparatively lower than with languages falling in the opposite category (Wheeler and Dillahunt, 2018) . To diminish the divide, we either need new methodologies that allow for the creation of language resources or the creation of a platform supporting the continuous development of such resources (Chakravarthi et al., 2018 (Chakravarthi et al., , 2019 (Chakravarthi et al., , 2020a Chakravarthi, 2020) .", |
| "cite_spans": [ |
| { |
| "start": 79, |
| "end": 94, |
| "text": "(Herring, 2008)", |
| "ref_id": "BIBREF19" |
| }, |
| { |
| "start": 188, |
| "end": 203, |
| "text": "(Herring, 2008)", |
| "ref_id": "BIBREF19" |
| }, |
| { |
| "start": 272, |
| "end": 295, |
| "text": "(Besacier et al., 2014)", |
| "ref_id": "BIBREF4" |
| }, |
| { |
| "start": 494, |
| "end": 509, |
| "text": "(Krauwer, 2003)", |
| "ref_id": "BIBREF23" |
| }, |
| { |
| "start": 610, |
| "end": 628, |
| "text": "(Warschauer, 2002)", |
| "ref_id": "BIBREF41" |
| }, |
| { |
| "start": 1092, |
| "end": 1121, |
| "text": "(Wheeler and Dillahunt, 2018)", |
| "ref_id": "BIBREF42" |
| }, |
| { |
| "start": 1317, |
| "end": 1343, |
| "text": "(Chakravarthi et al., 2018", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 1344, |
| "end": 1372, |
| "text": "(Chakravarthi et al., , 2019", |
| "ref_id": "BIBREF9" |
| }, |
| { |
| "start": 1373, |
| "end": 1402, |
| "text": "(Chakravarthi et al., , 2020a", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 1403, |
| "end": 1422, |
| "text": "Chakravarthi, 2020)", |
| "ref_id": "BIBREF7" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 631, |
| "end": 639, |
| "text": "Figure 1", |
| "ref_id": "FIGREF0" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "A person communicates their need for knowledge through words with a precise meaning, however from a machine point of view meaning( as well as specific acceptations) do not automatically come in conjunction with the word. Lexical-semantic resources come into play to bridge this gap, associating meanings to words. The most commonly used lexicalsemantic resource is the Princeton Word-Net(PWN) (Miller, 1998) . PWN is a lexical database for English that is organized around synsets. The popularity of PWN has spread in the language community and has lead to the generation of similar resources. One such attempt was made in India, a Country both rich in culture and languages. IndoWordNet(IWN) is first multilingual WordNet for Indian languages developed through the joint efforts of different reputed universities across the country. IWN developed intending to sketch India's cultures in length and breadth by including 18 languages out of 22 official languages (Dash et al., 2017) . All the languages in the IWN does not have the same synset IDs as PWN. For each synset in IWN, there is a corresponding synset in English that represent the same concept. In the paper, (Nair et al., 2019) authors extended this idea to align the IWN with PWN with the help of a large-scale multilingual resource called the Universal Knowledge Core(UKC) (Tawfik et al., 2014) . The UKC has a similar structure of PWN and is designed as a multilayered ontology that has a language-independent semantic layer called the Concept Core (CC) and a language-specific lexico-semantic layer called the Language Core (LC). Hence, in the UKC, the meaning of the words are represented not only by the synsets but also using lexical concepts (Giunchiglia et al., 2018) . Many works have been conducted under the UKC. Recent one is (Bella et al., 2020) , where authors explores the unique grammatical properties of natural language text and perform experiments on tokenisation, part of speech tagging and named entity recognition over real-world structured data. Another important one is , which introduces a large-scale lexical database that provides words of common origin and meaning across languages.", |
| "cite_spans": [ |
| { |
| "start": 393, |
| "end": 407, |
| "text": "(Miller, 1998)", |
| "ref_id": "BIBREF26" |
| }, |
| { |
| "start": 962, |
| "end": 981, |
| "text": "(Dash et al., 2017)", |
| "ref_id": "BIBREF13" |
| }, |
| { |
| "start": 1169, |
| "end": 1188, |
| "text": "(Nair et al., 2019)", |
| "ref_id": "BIBREF28" |
| }, |
| { |
| "start": 1336, |
| "end": 1357, |
| "text": "(Tawfik et al., 2014)", |
| "ref_id": "BIBREF39" |
| }, |
| { |
| "start": 1711, |
| "end": 1737, |
| "text": "(Giunchiglia et al., 2018)", |
| "ref_id": "BIBREF17" |
| }, |
| { |
| "start": 1800, |
| "end": 1820, |
| "text": "(Bella et al., 2020)", |
| "ref_id": "BIBREF3" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Before enriching an existing resource, ensuring the quality and finding hidden errors is recommended in order to avoid producing a low-quality extensive resource. Inadequate quality resources are not suitable to train artificial intelligence tools (Chakravarthi et al., 2020b) . One of the methods to ensure the quality of a resource is getting it validated using linguistic experts. Hiring linguists are expensive, or for some languages it is difficult. And another drawback is that contribution from one expert could be biased (Bonvillain, 2019) . Hence we need faster and cheaper ways to estimate the quality of a WordNet. Also, the main question here is what are the factors that define the quality of a WordNet. As per the literature, a WordNet development team usually does structural check and whether the WordNet can perform applications like information retrieval (Baeza-Yates et al., 2015) and question-answering (Moldovan and Rus, 2001) .", |
| "cite_spans": [ |
| { |
| "start": 248, |
| "end": 276, |
| "text": "(Chakravarthi et al., 2020b)", |
| "ref_id": "BIBREF11" |
| }, |
| { |
| "start": 529, |
| "end": 547, |
| "text": "(Bonvillain, 2019)", |
| "ref_id": "BIBREF5" |
| }, |
| { |
| "start": 923, |
| "end": 947, |
| "text": "(Moldovan and Rus, 2001)", |
| "ref_id": "BIBREF27" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "In this paper, we classify possible errors that affect the quality of the WordNet: schema errors and semantic errors. We list five schema errors from the IWN. We estimate semantic error by computing cosine similarity (Rahutomo et al., 2012) between two WordNets and within the WordNet using a pre-trained machine learning model (Zhou, 2016) for Indian languages. Using our approach, we generate a candidate set of synsets with very high semantic error that will be verified by native speakers in a This paper is structured as follows: Section 2 covers details about the language Malayalam and Malayam Wordnet (MWN), section 3 defines the quality of the WordNet and lists out the possible errors in WordNet, section 4 covers the procedure we used to find and establish the quality of the MWN, section 5 lists results and finally section 6 provides conclusions and direction for further work.", |
| "cite_spans": [ |
| { |
| "start": 217, |
| "end": 240, |
| "text": "(Rahutomo et al., 2012)", |
| "ref_id": "BIBREF33" |
| }, |
| { |
| "start": 328, |
| "end": 340, |
| "text": "(Zhou, 2016)", |
| "ref_id": "BIBREF43" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Malayalam (Asher, 2013 ) is a Dravidian language spoken in the Indian state of Kerala and union territories of Lakshadweep and Puducherry. Malayalam evolved from Tamil language after 16 century CE by Father of Malayalam Thunchaththu Ramanujan Ezhuthachan. In the early 21st century, Malayalam was spoken by around 38 million people all over the world (Asher, 2013) . In the past, media like television, newspaper, short stories, novels, etc. helped with the diffusion and prospering of Malayalam language and literature (Palakeel, 1996) . Nowadays smartphones, laptops, and Internet usage have become an important part of people's lives. The use of Malayalam language through these platforms has helped people to communicate, learn and express. Introduction of 'Unicode' technology in the Malayalam language helped correct spelling mistakes, grammatical errors, etc (Sooraj et al., 2018) (Santhosh et al., 2002) , which was initially possible only in the English language. The stories and articles were in the regional language, though they used English words for dif-ferent technical terms (Sasidharan et al., 2009) . For example, the Malayalam word for mobile phones is \"\u0d26\u0d42\u0d30\u0d2d\u0d3e\u0d37\u0d23\u0d36\u0d30\u0d4d \u0d35\u0d23\u0d38\u0d39\u0d3e\u0d2f\u0d3f\" (Dhoora bashana sravana sahayi), which was never used in any of the articles/blogs/news videos. The word 'Switch' in Malayalam was \"\u0d48\u0d35\u0d26\u0d41\u0d2f\u0d4d \u0d24 \u0d06\u0d17\u0d2e\u0d28 \u0d28\u0d3f\u0d2f\u0d28\u0d4d\u0d24\u0d30\u0d4d \u0d23 \u0d2f\u0d28\u0d4d\u0d24\u0d30\u0d4d \u0d02\" (Vaidhyutha agamana niyanthrana yanthram). But due to its complexity, these words were never used. The Malayalam language lost lots of words due to this reason(Prasanth, 2016). The Malayalam vocabulary is rich in terms from specific domains, with the notable example of terms reflecting family relations (George, 1972) . In Kerala, where Malayalam is mother tongue, family relations are essential and there are particular names for each type of relation. A notable example of this richness is represented by the English word cousin: \"the child of your aunt or uncle\". In Malayalam the term used to denote a cousin will change depending on age, gender, whether the relation is from the paternal or maternal side. Sometimes the background of the family is also a factor influencing the selection of a different term. There is no such thing as a direct translation of the word cousin in Malayalam.", |
| "cite_spans": [ |
| { |
| "start": 10, |
| "end": 22, |
| "text": "(Asher, 2013", |
| "ref_id": null |
| }, |
| { |
| "start": 351, |
| "end": 364, |
| "text": "(Asher, 2013)", |
| "ref_id": null |
| }, |
| { |
| "start": 520, |
| "end": 536, |
| "text": "(Palakeel, 1996)", |
| "ref_id": "BIBREF29" |
| }, |
| { |
| "start": 866, |
| "end": 887, |
| "text": "(Sooraj et al., 2018)", |
| "ref_id": "BIBREF38" |
| }, |
| { |
| "start": 888, |
| "end": 911, |
| "text": "(Santhosh et al., 2002)", |
| "ref_id": "BIBREF36" |
| }, |
| { |
| "start": 1091, |
| "end": 1116, |
| "text": "(Sasidharan et al., 2009)", |
| "ref_id": "BIBREF37" |
| }, |
| { |
| "start": 1664, |
| "end": 1678, |
| "text": "(George, 1972)", |
| "ref_id": "BIBREF15" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Malayalam Language", |
| "sec_num": "2" |
| }, |
| { |
| "text": "The Indian University, Amrita Vishwa Vidyapeetham, took part in the development of MWN (Rajendran and Soman, 2017) in 2011 as part of the project entitled \"Development of Dravidian wordnet: an integrated wordnet for Telugu, Tamil, Kannada and Malayalam\" which later integrated into IWN along with other Indian languages. The development of MWN is motivated by PWN. The initial idea of PWN was to aid the search in dictionaries, conceptually rather than alphabetically. Based on inspiration coming from current psycholinguistics (Foss, 1972) theories of human memory and its working, PWN organizes English nouns, verbs,adverbs, and adjectives into synonym sets or synset, each representing one underlying lexical concept. Synonyms are defined as words holding the same meaning in the same language (Miller, 1998) . A set of synonymous words is called synset. For instance, \"head\" and \"caput\" are the synsets with the meaning \"the front or the upper part of the body of animals bearing the face and brains\". In PWN, synsets are associated with a gloss. A Gloss defines the meaning of a synset, the concept associated with the synset. MWN uses the same structure and lists the following entries for each concept: synset ID, synsets, gloss and example sentence. Synset ids used in MWN are different from PWN as the expand approach was used to develop MWN. In the expand approach, Hindi synsets are translated into Malayalam, this ensures concepts which are culturally relevant to India are retained as part of MWN. In our study, we are using source files of MWN we received from Amrita Vishwa Vidyapeetham. There are 30139 synsets available in MWN. In which, 20071 synsets are nouns, 3311 synsets are verbs, 501 synsets are adverbs, and 6256 synsets are adjectives.", |
| "cite_spans": [ |
| { |
| "start": 87, |
| "end": 114, |
| "text": "(Rajendran and Soman, 2017)", |
| "ref_id": "BIBREF34" |
| }, |
| { |
| "start": 528, |
| "end": 540, |
| "text": "(Foss, 1972)", |
| "ref_id": null |
| }, |
| { |
| "start": 797, |
| "end": 811, |
| "text": "(Miller, 1998)", |
| "ref_id": "BIBREF26" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Malayalam Language", |
| "sec_num": "2" |
| }, |
| { |
| "text": "Within the translation industry, three terms are used somewhat interchangeably to refer to quality-related initiatives: quality assessment (House, 2014) , quality assurance (V\u00eelceanu, 2017) and quality control (Ibba and S\u00f6ll, 1999 ).", |
| "cite_spans": [ |
| { |
| "start": 139, |
| "end": 152, |
| "text": "(House, 2014)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 173, |
| "end": 189, |
| "text": "(V\u00eelceanu, 2017)", |
| "ref_id": "BIBREF40" |
| }, |
| { |
| "start": 210, |
| "end": 230, |
| "text": "(Ibba and S\u00f6ll, 1999", |
| "ref_id": "BIBREF21" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Defining Quality", |
| "sec_num": "3" |
| }, |
| { |
| "text": "\u2022 Quality assessment or quality evaluation is the measurement of the extent to which a product complies with quality specifications (Lommel et al., 2014 ).", |
| "cite_spans": [ |
| { |
| "start": 132, |
| "end": 152, |
| "text": "(Lommel et al., 2014", |
| "ref_id": "BIBREF24" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Defining Quality", |
| "sec_num": "3" |
| }, |
| { |
| "text": "\u2022 Quality assurance refers to ways of preventing mistakes or defects in manufactured products and avoiding problems when delivering solutions or services to customers. Quality assurance relies on continual assessment of quality (Lommel et al., 2014) .", |
| "cite_spans": [ |
| { |
| "start": 228, |
| "end": 249, |
| "text": "(Lommel et al., 2014)", |
| "ref_id": "BIBREF24" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Defining Quality", |
| "sec_num": "3" |
| }, |
| { |
| "text": "\u2022 Quality control is the process of checking whether manufactured products meet stated quality specifications. While quality assurance relates to how a process is performed or how a product is made, quality control is tied more to the inspection aspect of quality management (Lommel et al., 2014) .", |
| "cite_spans": [ |
| { |
| "start": 275, |
| "end": 296, |
| "text": "(Lommel et al., 2014)", |
| "ref_id": "BIBREF24" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Defining Quality", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Based on the definition, quality control requires specifications (Gouadec, 2010) which can then be used to assess whether a resource meets quality thresholds. In our work, we are performing quality control of the MWN. We are defining the minimum specifications in order to evaluate the WordNet and assess whether it is a good quality resource.", |
| "cite_spans": [ |
| { |
| "start": 65, |
| "end": 80, |
| "text": "(Gouadec, 2010)", |
| "ref_id": "BIBREF18" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Defining Quality", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Multidimensional Quality Metrics (MQM) (Lommel et al., 2014) provides a framework for describing and defining quality metrics used to assess the quality of translated texts and to identify specific issues in those texts. The scope of this framework is limited to quality assessment of translated content. It does not apply to the evaluation of the translation process or projects. MQM defines the quality as adherence of the text to appropriate specifications. As per MQM, specifications are a description of requirements for the translation. In our work, we have been inspired to adapt ideas used as part of MQM for defining quality specifications of the WordNet. In MQM, a target document quality score is defined as T Q = 100 \u2212 T P + SP TQ being the quality score, TP the penalties for target content, and SP the penalty tied to source content. We then define a resource as being of good quality if the resource is free from errors.", |
| "cite_spans": [ |
| { |
| "start": 39, |
| "end": 60, |
| "text": "(Lommel et al., 2014)", |
| "ref_id": "BIBREF24" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Defining Quality", |
| "sec_num": "3" |
| }, |
| { |
| "text": "We define two categories of errors in Word-Net: schema errors and semantic errors.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Defining Quality", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Schema errors occur when there is a problem with the file's structure or order, or an invalid character is included. There are five observed schema errors. They are:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Schema Errors", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "\u2022 Category A: Empty gloss and/or synset and/or example sentence We consider a record as having unique synset id, gloss, example sentence and synset. Category A is the number of records with empty gloss and/or empty example sentence and/or empty synset (Figure 2) . In some WordNets, example sentences are optional, so we will do a manual check in the beginning to verify metadata. Category B represents the number of records having a poorly defined gloss ( Figure 3 ). Poorly defined glosses are very general, and computing this value is a bit challenging. Hence, we consider the number of records that have the same gloss as the corresponding synset. Category C is the number of records that have unknown characters in any of the fields (Figure 4 ). Category D is very rare but possible, including records that have poorly defined parts of speech ( Figure 5 ). Four parts of speech are included in MWN: noun, adjective, verb and adverb. Due to errors during the development phase, these values may have spelling mistakes. This eventually affects the quality of the WordNet. Category E includes the number of records having the same gloss ( Figure 6 ). For these records, synset id, example sentence and synsets are different, gloss is the same.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 252, |
| "end": 262, |
| "text": "(Figure 2)", |
| "ref_id": "FIGREF1" |
| }, |
| { |
| "start": 457, |
| "end": 466, |
| "text": "Figure 3", |
| "ref_id": null |
| }, |
| { |
| "start": 739, |
| "end": 749, |
| "text": "(Figure 4", |
| "ref_id": "FIGREF2" |
| }, |
| { |
| "start": 852, |
| "end": 860, |
| "text": "Figure 5", |
| "ref_id": null |
| }, |
| { |
| "start": 1143, |
| "end": 1151, |
| "text": "Figure 6", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Schema Errors", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "Meanings can be represented in many languages if the corresponding concept is available across all languages that are taken into account (Mercer, 2002) . Meaning is defined as what is meant by a word, text, concept, or action. For example, the word \"rain\" means \"water falling in drops from vapor condensed in the atmosphere\" and the same meaning is represented in Italian as \"pioggia\", in Dutch as \"regen\". It is also worth keeping in mind that the word \"rain\" has two additional possible meanings, available depending on the context the word is used in.", |
| "cite_spans": [ |
| { |
| "start": 137, |
| "end": 151, |
| "text": "(Mercer, 2002)", |
| "ref_id": "BIBREF25" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Semantic Errors", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "Semantic errors are estimated when comparing across WordNets; checking how good conceptual alignment between the languages is. In our work, we are using PWN and HWN to find the semantic error in MWN. Semantic errors are caused by factors like cultural differences, linguistic differences, and mistakes made by humans. Figure 7 listing out some examples from these category and we have taken these values from IWN. Figures 7a, 7c, 7d and 7e are synsets specific to the language and culture. Finding out this candidate set could help the researchers to understand the diversity and similarity (Giunchiglia et al., 2017) across languages.", |
| "cite_spans": [ |
| { |
| "start": 591, |
| "end": 617, |
| "text": "(Giunchiglia et al., 2017)", |
| "ref_id": "BIBREF16" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 318, |
| "end": 326, |
| "text": "Figure 7", |
| "ref_id": "FIGREF4" |
| }, |
| { |
| "start": 414, |
| "end": 432, |
| "text": "Figures 7a, 7c, 7d", |
| "ref_id": "FIGREF4" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Semantic Errors", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "In this section, we evaluate schema errors and semantic errors for the MWN. We estimated the schema errors of the resource as listed in the table 1. There are three categories of schema errors presented in the MWN; category A, B and E. 181 records have empty example sentence, 1789 records have empty synsets and 2374 records have poorly defined glosses. And then we found 925 records with duplicate gloss.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Evaluating MWN", |
| "sec_num": "4" |
| }, |
| { |
| "text": "We sampled synsets common to English, Figure 6: Example for duplicate gloss (a) Synset \"temple\" vs \"\u0d36\u0d3f\u0d35 _\u0d15\u0d4d\u0d47\u0d37\u0d24\u0d30\u0d4d \u0d02\"(shiva_kshetram) (b) Synset \"site\" vs \"\u0d05\u0d36\u0d41\u0d26\u0d4d\u0d27\u0d38\u0d4d\u0d25\u0d32\u0d02\"(ashudhasthalam) (c) Synset \"expert\" vs \"\u0d28\u0d35\u0d30\u0d24\u0d4d\u0d28\u0d02\" (navarathnam) (d) Synset \"dish\" vs \"\u0d38\u0d41\u0d16\u0d3f\u0d2f\u0d28\u0d4d \"(sukhiyan (e) Synset \"rice\" vs \"\u0d2c\u0d38\u0d4d\u0d2e\u0d24\u0d3f _\u0d05\u0d30\u0d3f\"(basmati _ari) Hindi and Malayalam WordNets; this yielded 18938 synsets. We then computed cosine similarity in order to identify synsets tied to poor semantic alignment and through this be able to identify semantic errors. We propose a cosine similarity score based classification parsed into three classes: low alignment for scores lower than 0.4; moderate for scores between 0.4 and 0.7; strong for scores greater than 0.7. We have used the threshold of minimum 0.4 inspired by the work by (Khodak et al., 2017) . In (Khodak et al., 2017) authors automatically generate WordNet data by using a machine-readable dictionary. Based on our classification, we will consider the sets of synsets having cosine similarity scores below 0.4 as requiring attention.", |
| "cite_spans": [ |
| { |
| "start": 797, |
| "end": 818, |
| "text": "(Khodak et al., 2017)", |
| "ref_id": "BIBREF22" |
| }, |
| { |
| "start": 824, |
| "end": 845, |
| "text": "(Khodak et al., 2017)", |
| "ref_id": "BIBREF22" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Evaluating MWN", |
| "sec_num": "4" |
| }, |
| { |
| "text": "We used two approaches to measure semantic errors: we computed cosine similarity(semantic alignment) between two languages; we then calculated semantic alignment within the language. Algorithm 1 shows the steps followed in computing the semantic similarity between English gloss and Malayalam gloss. We have used the same steps to calculate the semantic similarity between English word sense and Malayalam word sense. And repeat the above steps also for language Hindi. Algorithm 2 shows the steps followed in computing the semantic similarity between a word sense in the synset and the gloss of MWN.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Evaluating MWN", |
| "sec_num": "4" |
| }, |
| { |
| "text": "A pre-trained machine learning model was used in order to compute semantic alignment and cosine similarity. There are a number of pre-trained models available for similarity checking. Multilingual BERT(mBERT) (Pires et al., 2019) , XLM-RoBERTa (Conneau et al., 2019) and Sentence transformers (Reimers and Gurevych, 2019) are only few of them. Multilingual BERT(mBERT) and XLM-RoBERTa have been known to produce less than ideal sentence representations when deployed out-ofbox. They would additionally pose the problem of coming with non-aligned vector spaces across languages. As a result, sentences with the same semantic content, expressed with different languages, would be mapped to different vector spaces. As part of our work we used sentence transformers: a Python framework for state of the art sentence and text embeddings that can be used to compute sentence/text embeddings for more than 100 languages. These embeddings can be compared through cosine similarity, allowing for the identification of sentences with similar meaning. The framework is based on PyTorch (Paszke et al., 2019) , Transformers and offers a large collection of pre-trained models tuned for various tasks. We have used the pre-trained multilingual model stsb-xlm-r-multilingual and aligned vector spaces allowing for similar inputs across languages to be mapped close within the same vector space. XLM-R supports 100 languages including 13 Indian languages, and is as such able to handle linguistic inputs without the need to specify what the input language is upfront. The model produces similar embeddings as the bert-base-nli-stsb-mean-token model. ", |
| "cite_spans": [ |
| { |
| "start": 209, |
| "end": 229, |
| "text": "(Pires et al., 2019)", |
| "ref_id": "BIBREF31" |
| }, |
| { |
| "start": 244, |
| "end": 266, |
| "text": "(Conneau et al., 2019)", |
| "ref_id": "BIBREF12" |
| }, |
| { |
| "start": 293, |
| "end": 321, |
| "text": "(Reimers and Gurevych, 2019)", |
| "ref_id": "BIBREF35" |
| }, |
| { |
| "start": 1076, |
| "end": 1097, |
| "text": "(Paszke et al., 2019)", |
| "ref_id": "BIBREF30" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Evaluating MWN", |
| "sec_num": "4" |
| }, |
| { |
| "text": "There are 30139 Synsets available in the Malayalam language. Using our approach, we estimated that 14% of the Synsets have schema errors. Figure 8 summarizes cosine similarity score distribution. 30% of gloss-wise similarity scores for PWN-MWN synsets fall in the low alignment category, the score goes down to 17% if looking at word sense level alignment for PWN-MWN entries. Interestingly, when looking at HWN-MWN data, synsets falling into the low alignment category decrease to 4% when looking at gloss-wise alignment; 2% when looking at word sense-wise alignment. Looking at within language data, reflects that 20% of MWN word senses are tied to low scores(below 0.4). Figure 9 shows the distribution of similarity agreement within the Malayalam language and Figure 10 shows the similarity distribution of Malayalam language between English and Hindi.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 138, |
| "end": 146, |
| "text": "Figure 8", |
| "ref_id": "FIGREF5" |
| }, |
| { |
| "start": 674, |
| "end": 683, |
| "text": "Figure 9", |
| "ref_id": "FIGREF6" |
| }, |
| { |
| "start": 765, |
| "end": 774, |
| "text": "Figure 10", |
| "ref_id": "FIGREF0" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "Computing semantic similarity within the language helps us understand and isolate poorly aligned word senses and better word senses. Results aligned how Malayalam language entries are more semantically aligned with Hindi than English, which is expected.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "This paper proposes specifications to be considered for a good quality lexical-semantic resource. We believe a resource without schema and semantic errors to be a useful resource. We defined schema errors and semantic errors for WordNet and estimated the presence of both for MWN. 14% of the Malayalam resource has schema errors. An average of 23.5% meanings representing the same concepts across English and Malayalam are tied to low similarity scores( poor semantic alignment). Only 3% of meanings representing the same concepts across Hindi and Malayalam have low similarity scores. Our approach concluded that when looking at within language data, 20% of MWN word senses are poorly aligned with their gloss. With this finding, we can classify MWN into two parts: Synsets requiring human revision and Synsets that can be readily shared with the world. We will be using MWN as part of a web application that supports end-Users in finding translations.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion and Future work", |
| "sec_num": "6" |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": "We are not taking these values as the final deciding factor. We will be using this low score synsets as a candidate set for our crowdsourcing application. This application will have different tasks like define gloss, provide Synset, validate the gloss, and so on.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "acknowledgement", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": { |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "CASSA: A context-aware synonym simplification algorithm", |
| "authors": [ |
| { |
| "first": "Ricardo", |
| "middle": [], |
| "last": "Baeza-Yates", |
| "suffix": "" |
| }, |
| { |
| "first": "Luz", |
| "middle": [], |
| "last": "Rello", |
| "suffix": "" |
| }, |
| { |
| "first": "Julia", |
| "middle": [], |
| "last": "Dembowski", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "Proceedings of the 2015 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies", |
| "volume": "", |
| "issue": "", |
| "pages": "1380--1385", |
| "other_ids": { |
| "DOI": [ |
| "10.3115/v1/N15-1156" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ricardo Baeza-Yates, Luz Rello, and Julia Dem- bowski. 2015. CASSA: A context-aware syn- onym simplification algorithm. In Proceedings of the 2015 Conference of the North Ameri- can Chapter of the Association for Computa- tional Linguistics: Human Language Technolo- gies, pages 1380-1385, Denver, Colorado. Asso- ciation for Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "Cognet: A large-scale cognate database", |
| "authors": [ |
| { |
| "first": "Khuyagbaatar", |
| "middle": [], |
| "last": "Batsuren", |
| "suffix": "" |
| }, |
| { |
| "first": "G\u00e1bor", |
| "middle": [], |
| "last": "Bella", |
| "suffix": "" |
| }, |
| { |
| "first": "Fausto", |
| "middle": [], |
| "last": "Giunchiglia", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "3136--3145", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Khuyagbaatar Batsuren, G\u00e1bor Bella, and Fausto Giunchiglia. 2019. Cognet: A large-scale cog- nate database. In Proceedings of the 57th An- nual Meeting of the Association for Computa- tional Linguistics, pages 3136-3145.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Exploring the language of data", |
| "authors": [ |
| { |
| "first": "G\u00e1bor", |
| "middle": [], |
| "last": "Bella", |
| "suffix": "" |
| }, |
| { |
| "first": "Linda", |
| "middle": [], |
| "last": "Gremes", |
| "suffix": "" |
| }, |
| { |
| "first": "Fausto", |
| "middle": [], |
| "last": "Giunchiglia", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Proceedings of the 28th International Conference on Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "6638--6648", |
| "other_ids": { |
| "DOI": [ |
| "10.18653/v1/2020.coling-main.582" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "G\u00e1bor Bella, Linda Gremes, and Fausto Giunchiglia. 2020. Exploring the language of data. In Proceedings of the 28th Interna- tional Conference on Computational Linguistics, pages 6638-6648, Barcelona, Spain (Online). International Committee on Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "Automatic speech recognition for under-resourced languages: A survey", |
| "authors": [ |
| { |
| "first": "Laurent", |
| "middle": [], |
| "last": "Besacier", |
| "suffix": "" |
| }, |
| { |
| "first": "Etienne", |
| "middle": [], |
| "last": "Barnard", |
| "suffix": "" |
| }, |
| { |
| "first": "Alexey", |
| "middle": [], |
| "last": "Karpov", |
| "suffix": "" |
| }, |
| { |
| "first": "Tanja", |
| "middle": [], |
| "last": "Schultz", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "Speech Communication", |
| "volume": "56", |
| "issue": "", |
| "pages": "85--100", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Laurent Besacier, Etienne Barnard, Alexey Kar- pov, and Tanja Schultz. 2014. Automatic speech recognition for under-resourced languages: A survey. Speech Communication, 56:85-100.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Language, culture, and communication: The meaning of messages. Rowman & Littlefield", |
| "authors": [ |
| { |
| "first": "Nancy", |
| "middle": [], |
| "last": "Bonvillain", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nancy Bonvillain. 2019. Language, culture, and communication: The meaning of messages. Row- man & Littlefield.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Crowdsourcing", |
| "authors": [ |
| { |
| "first": "C", |
| "middle": [], |
| "last": "Daren", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Brabham", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Daren C Brabham. 2013. Crowdsourcing. Mit Press.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Leveraging orthographic information to improve machine translation of under-resourced languages", |
| "authors": [ |
| { |
| "first": "Chakravarthi", |
| "middle": [], |
| "last": "Bharathi Raja", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bharathi Raja Chakravarthi. 2020. Leveraging orthographic information to improve machine translation of under-resourced languages. Ph.D. thesis, NUI Galway.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Improving wordnets for under-resourced languages using machine translation", |
| "authors": [ |
| { |
| "first": "Mihael", |
| "middle": [], |
| "last": "Bharathi Raja Chakravarthi", |
| "suffix": "" |
| }, |
| { |
| "first": "John", |
| "middle": [ |
| "P" |
| ], |
| "last": "Arcan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Mccrae", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "Proceedings of the 9th Global WordNet Conference", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bharathi Raja Chakravarthi, Mihael Arcan, and John P McCrae. 2018. Improving wordnets for under-resourced languages using machine trans- lation. In Proceedings of the 9th Global WordNet Conference (GWC 2018), page 78.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "WordNet gloss translation for under-resourced languages using multilingual neural machine translation", |
| "authors": [ |
| { |
| "first": "Mihael", |
| "middle": [], |
| "last": "Bharathi Raja Chakravarthi", |
| "suffix": "" |
| }, |
| { |
| "first": "John", |
| "middle": [ |
| "P" |
| ], |
| "last": "Arcan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Mccrae", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Proceedings of the Second Workshop on Multilingualism at the Intersection of Knowledge Bases and Machine Translation", |
| "volume": "", |
| "issue": "", |
| "pages": "1--7", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bharathi Raja Chakravarthi, Mihael Arcan, and John P. McCrae. 2019. WordNet gloss transla- tion for under-resourced languages using multi- lingual neural machine translation. In Proceed- ings of the Second Workshop on Multilingualism at the Intersection of Knowledge Bases and Ma- chine Translation, pages 1-7, Dublin, Ireland. European Association for Machine Translation.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "Bilingual lexicon induction across orthographically-distinct under-resourced Dravidian languages", |
| "authors": [ |
| { |
| "first": "Navaneethan", |
| "middle": [], |
| "last": "Bharathi Raja Chakravarthi", |
| "suffix": "" |
| }, |
| { |
| "first": "Mihael", |
| "middle": [], |
| "last": "Rajasekaran", |
| "suffix": "" |
| }, |
| { |
| "first": "Kevin", |
| "middle": [], |
| "last": "Arcan", |
| "suffix": "" |
| }, |
| { |
| "first": "Noel", |
| "middle": [ |
| "E" |
| ], |
| "last": "Mcguinness", |
| "suffix": "" |
| }, |
| { |
| "first": "John", |
| "middle": [ |
| "P" |
| ], |
| "last": "O'connor", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Mccrae", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Proceedings of the 7th Workshop on NLP for Similar Languages, Varieties and Dialects", |
| "volume": "", |
| "issue": "", |
| "pages": "57--69", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bharathi Raja Chakravarthi, Navaneethan Ra- jasekaran, Mihael Arcan, Kevin McGuinness, Noel E. O'Connor, and John P. McCrae. 2020a. Bilingual lexicon induction across orthographically-distinct under-resourced Dra- vidian languages. In Proceedings of the 7th Workshop on NLP for Similar Languages, Va- rieties and Dialects, pages 57-69, Barcelona, Spain (Online). International Committee on Computational Linguistics (ICCL).", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "A survey of orthographic information in machine translation", |
| "authors": [ |
| { |
| "first": "Priya", |
| "middle": [], |
| "last": "Bharathi Raja Chakravarthi", |
| "suffix": "" |
| }, |
| { |
| "first": "Mihael", |
| "middle": [], |
| "last": "Rani", |
| "suffix": "" |
| }, |
| { |
| "first": "John", |
| "middle": [ |
| "P" |
| ], |
| "last": "Arcan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Mccrae", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:2008.01391" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bharathi Raja Chakravarthi, Priya Rani, Mihael Arcan, and John P McCrae. 2020b. A survey of orthographic information in machine translation. arXiv preprint arXiv:2008.01391.", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "Unsupervised cross-lingual representation learning at scale", |
| "authors": [ |
| { |
| "first": "Alexis", |
| "middle": [], |
| "last": "Conneau", |
| "suffix": "" |
| }, |
| { |
| "first": "Kartikay", |
| "middle": [], |
| "last": "Khandelwal", |
| "suffix": "" |
| }, |
| { |
| "first": "Naman", |
| "middle": [], |
| "last": "Goyal", |
| "suffix": "" |
| }, |
| { |
| "first": "Vishrav", |
| "middle": [], |
| "last": "Chaudhary", |
| "suffix": "" |
| }, |
| { |
| "first": "Guillaume", |
| "middle": [], |
| "last": "Wenzek", |
| "suffix": "" |
| }, |
| { |
| "first": "Francisco", |
| "middle": [], |
| "last": "Guzm\u00e1n", |
| "suffix": "" |
| }, |
| { |
| "first": "Edouard", |
| "middle": [], |
| "last": "Grave", |
| "suffix": "" |
| }, |
| { |
| "first": "Myle", |
| "middle": [], |
| "last": "Ott", |
| "suffix": "" |
| }, |
| { |
| "first": "Luke", |
| "middle": [], |
| "last": "Zettlemoyer", |
| "suffix": "" |
| }, |
| { |
| "first": "Veselin", |
| "middle": [], |
| "last": "Stoyanov", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1911.02116" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzm\u00e1n, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Unsupervised cross-lingual representation learn- ing at scale. arXiv preprint arXiv:1911.02116.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "The WordNet in Indian Languages", |
| "authors": [ |
| { |
| "first": "Pushpak", |
| "middle": [], |
| "last": "Niladri Sekhar Dash", |
| "suffix": "" |
| }, |
| { |
| "first": "Jyoti", |
| "middle": [ |
| "D" |
| ], |
| "last": "Bhattacharyya", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Pawar", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Niladri Sekhar Dash, Pushpak Bhattacharyya, and Jyoti D Pawar. 2017. The WordNet in Indian Languages. Springer.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "Western influence on Malayalam language and literature", |
| "authors": [ |
| { |
| "first": "George", |
| "middle": [], |
| "last": "Karimpumannil Mathai", |
| "suffix": "" |
| } |
| ], |
| "year": 1972, |
| "venue": "Sahitya Akademi", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Karimpumannil Mathai George. 1972. Western in- fluence on Malayalam language and literature. Sahitya Akademi.", |
| "links": null |
| }, |
| "BIBREF16": { |
| "ref_id": "b16", |
| "title": "Understanding and exploiting language diversity", |
| "authors": [ |
| { |
| "first": "Fausto", |
| "middle": [], |
| "last": "Giunchiglia", |
| "suffix": "" |
| }, |
| { |
| "first": "Khuyagbaatar", |
| "middle": [], |
| "last": "Batsuren", |
| "suffix": "" |
| }, |
| { |
| "first": "Gabor", |
| "middle": [], |
| "last": "Bella", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "IJCAI", |
| "volume": "", |
| "issue": "", |
| "pages": "4009--4017", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Fausto Giunchiglia, Khuyagbaatar Batsuren, and Gabor Bella. 2017. Understanding and ex- ploiting language diversity. In IJCAI, pages 4009-4017.", |
| "links": null |
| }, |
| "BIBREF17": { |
| "ref_id": "b17", |
| "title": "One world-seven thousand languages", |
| "authors": [ |
| { |
| "first": "Fausto", |
| "middle": [], |
| "last": "Giunchiglia", |
| "suffix": "" |
| }, |
| { |
| "first": "Khuyagbaatar", |
| "middle": [], |
| "last": "Batsuren", |
| "suffix": "" |
| }, |
| { |
| "first": "Abed Alhakim", |
| "middle": [], |
| "last": "Freihat", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "Proceedings 19th International Conference on Computational Linguistics and Intelligent Text Processing", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Fausto Giunchiglia, Khuyagbaatar Batsuren, and Abed Alhakim Freihat. 2018. One world-seven thousand languages. In Proceedings 19th In- ternational Conference on Computational Lin- guistics and Intelligent Text Processing, Ci- Cling2018, 18-24 March 2018.", |
| "links": null |
| }, |
| "BIBREF18": { |
| "ref_id": "b18", |
| "title": "Quality in translation. Handbook of translation studies", |
| "authors": [ |
| { |
| "first": "Daniel", |
| "middle": [], |
| "last": "Gouadec", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "", |
| "volume": "1", |
| "issue": "", |
| "pages": "270--275", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Daniel Gouadec. 2010. Quality in translation. Handbook of translation studies, 1:270-275.", |
| "links": null |
| }, |
| "BIBREF19": { |
| "ref_id": "b19", |
| "title": "Language and the internet. The International Encyclopedia of Communication", |
| "authors": [ |
| { |
| "first": "C", |
| "middle": [], |
| "last": "Susan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Herring", |
| "suffix": "" |
| } |
| ], |
| "year": 2008, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Susan C Herring. 2008. Language and the internet. The International Encyclopedia of Communica- tion.", |
| "links": null |
| }, |
| "BIBREF20": { |
| "ref_id": "b20", |
| "title": "Translation quality assessment: Past and present", |
| "authors": [ |
| { |
| "first": "Juliane", |
| "middle": [], |
| "last": "House", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "Translation: A multidisciplinary approach", |
| "volume": "", |
| "issue": "", |
| "pages": "241--264", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Juliane House. 2014. Translation quality as- sessment: Past and present. In Translation: A multidisciplinary approach, pages 241-264. Springer.", |
| "links": null |
| }, |
| "BIBREF21": { |
| "ref_id": "b21", |
| "title": "Quality control mechanisms during translation", |
| "authors": [ |
| { |
| "first": "Michael", |
| "middle": [], |
| "last": "Ibba", |
| "suffix": "" |
| }, |
| { |
| "first": "Dieter", |
| "middle": [], |
| "last": "S\u00f6ll", |
| "suffix": "" |
| } |
| ], |
| "year": 1999, |
| "venue": "Science", |
| "volume": "286", |
| "issue": "5446", |
| "pages": "1893--1897", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Michael Ibba and Dieter S\u00f6ll. 1999. Quality con- trol mechanisms during translation. Science, 286(5446):1893-1897.", |
| "links": null |
| }, |
| "BIBREF22": { |
| "ref_id": "b22", |
| "title": "Automated WordNet construction using word embeddings", |
| "authors": [ |
| { |
| "first": "Mikhail", |
| "middle": [], |
| "last": "Khodak", |
| "suffix": "" |
| }, |
| { |
| "first": "Andrej", |
| "middle": [], |
| "last": "Risteski", |
| "suffix": "" |
| }, |
| { |
| "first": "Christiane", |
| "middle": [], |
| "last": "Fellbaum", |
| "suffix": "" |
| }, |
| { |
| "first": "Sanjeev", |
| "middle": [], |
| "last": "Arora", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Proceedings of the 1st Workshop on Sense, Concept and Entity Representations and their Applications", |
| "volume": "", |
| "issue": "", |
| "pages": "12--23", |
| "other_ids": { |
| "DOI": [ |
| "10.18653/v1/W17-1902" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Mikhail Khodak, Andrej Risteski, Christiane Fell- baum, and Sanjeev Arora. 2017. Automated WordNet construction using word embeddings. In Proceedings of the 1st Workshop on Sense, Concept and Entity Representations and their Applications, pages 12-23, Valencia, Spain. As- sociation for Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF23": { |
| "ref_id": "b23", |
| "title": "The basic language resource kit (blark) as the first milestone for the language resources roadmap", |
| "authors": [ |
| { |
| "first": "Steven", |
| "middle": [], |
| "last": "Krauwer", |
| "suffix": "" |
| } |
| ], |
| "year": 2003, |
| "venue": "Proceedings of SPECOM 2003", |
| "volume": "", |
| "issue": "", |
| "pages": "8--15", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Steven Krauwer. 2003. The basic language re- source kit (blark) as the first milestone for the language resources roadmap. Proceedings of SPECOM 2003, pages 8-15.", |
| "links": null |
| }, |
| "BIBREF24": { |
| "ref_id": "b24", |
| "title": "Multidimensional quality metrics (mqm): A framework for declaring and describing translation quality metrics", |
| "authors": [ |
| { |
| "first": "Arle", |
| "middle": [], |
| "last": "Lommel", |
| "suffix": "" |
| }, |
| { |
| "first": "Hans", |
| "middle": [], |
| "last": "Uszkoreit", |
| "suffix": "" |
| }, |
| { |
| "first": "Aljoscha", |
| "middle": [], |
| "last": "Burchardt", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "Tradum\u00e0tica", |
| "volume": "", |
| "issue": "12", |
| "pages": "455--463", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Arle Lommel, Hans Uszkoreit, and Aljoscha Bur- chardt. 2014. Multidimensional quality metrics (mqm): A framework for declaring and describ- ing translation quality metrics. Tradum\u00e0tica, (12):0455-463.", |
| "links": null |
| }, |
| "BIBREF25": { |
| "ref_id": "b25", |
| "title": "Words and minds: How we use language to think together", |
| "authors": [ |
| { |
| "first": "Neil", |
| "middle": [], |
| "last": "Mercer", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Neil Mercer. 2002. Words and minds: How we use language to think together. Routledge.", |
| "links": null |
| }, |
| "BIBREF26": { |
| "ref_id": "b26", |
| "title": "WordNet: An electronic lexical database", |
| "authors": [ |
| { |
| "first": "A", |
| "middle": [], |
| "last": "George", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Miller", |
| "suffix": "" |
| } |
| ], |
| "year": 1998, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "George A Miller. 1998. WordNet: An electronic lexical database. MIT press.", |
| "links": null |
| }, |
| "BIBREF27": { |
| "ref_id": "b27", |
| "title": "Logic form transformation of WordNet and its applicability to question answering", |
| "authors": [ |
| { |
| "first": "Dan", |
| "middle": [], |
| "last": "Moldovan", |
| "suffix": "" |
| }, |
| { |
| "first": "Vasile", |
| "middle": [], |
| "last": "Rus", |
| "suffix": "" |
| } |
| ], |
| "year": 2001, |
| "venue": "Proceedings of the 39th Annual Meeting of the Association for Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "402--409", |
| "other_ids": { |
| "DOI": [ |
| "10.3115/1073012.1073064" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dan Moldovan and Vasile Rus. 2001. Logic form transformation of WordNet and its applicabil- ity to question answering. In Proceedings of the 39th Annual Meeting of the Association for Com- putational Linguistics, pages 402-409, Toulouse, France. Association for Computational Linguis- tics.", |
| "links": null |
| }, |
| "BIBREF28": { |
| "ref_id": "b28", |
| "title": "Aligning the indowordnet with the princeton wordnet", |
| "authors": [ |
| { |
| "first": "Rajendran", |
| "middle": [], |
| "last": "Nandu Chandran Nair", |
| "suffix": "" |
| }, |
| { |
| "first": "Khuyagbaatar", |
| "middle": [], |
| "last": "Sankara Velayuthan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Batsuren", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Proceedings of the 3rd International Conference on Natural Language and Speech Processing", |
| "volume": "", |
| "issue": "", |
| "pages": "9--16", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nandu Chandran Nair, Rajendran Sankara Ve- layuthan, and Khuyagbaatar Batsuren. 2019. Aligning the indowordnet with the princeton wordnet. In Proceedings of the 3rd International Conference on Natural Language and Speech Pro- cessing, pages 9-16.", |
| "links": null |
| }, |
| "BIBREF29": { |
| "ref_id": "b29", |
| "title": "Twentieth century malayalam literature. Handbook of Twentieth-century Literatures of India", |
| "authors": [ |
| { |
| "first": "Thomas", |
| "middle": [], |
| "last": "Palakeel", |
| "suffix": "" |
| } |
| ], |
| "year": 1996, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "180--200", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Thomas Palakeel. 1996. Twentieth century malay- alam literature. Handbook of Twentieth-century Literatures of India, pages 180-200.", |
| "links": null |
| }, |
| "BIBREF30": { |
| "ref_id": "b30", |
| "title": "Pytorch: An imperative style, high-performance deep learning library", |
| "authors": [ |
| { |
| "first": "Adam", |
| "middle": [], |
| "last": "Paszke", |
| "suffix": "" |
| }, |
| { |
| "first": "Sam", |
| "middle": [], |
| "last": "Gross", |
| "suffix": "" |
| }, |
| { |
| "first": "Francisco", |
| "middle": [], |
| "last": "Massa", |
| "suffix": "" |
| }, |
| { |
| "first": "Adam", |
| "middle": [], |
| "last": "Lerer", |
| "suffix": "" |
| }, |
| { |
| "first": "James", |
| "middle": [], |
| "last": "Bradbury", |
| "suffix": "" |
| }, |
| { |
| "first": "Gregory", |
| "middle": [], |
| "last": "Chanan", |
| "suffix": "" |
| }, |
| { |
| "first": "Trevor", |
| "middle": [], |
| "last": "Killeen", |
| "suffix": "" |
| }, |
| { |
| "first": "Zeming", |
| "middle": [], |
| "last": "Lin", |
| "suffix": "" |
| }, |
| { |
| "first": "Natalia", |
| "middle": [], |
| "last": "Gimelshein", |
| "suffix": "" |
| }, |
| { |
| "first": "Luca", |
| "middle": [], |
| "last": "Antiga", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Advances in neural information processing systems", |
| "volume": "", |
| "issue": "", |
| "pages": "8026--8037", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. 2019. Pytorch: An imper- ative style, high-performance deep learning li- brary. In Advances in neural information pro- cessing systems, pages 8026-8037.", |
| "links": null |
| }, |
| "BIBREF31": { |
| "ref_id": "b31", |
| "title": "How multilingual is multilingual bert? arXiv preprint", |
| "authors": [ |
| { |
| "first": "Telmo", |
| "middle": [], |
| "last": "Pires", |
| "suffix": "" |
| }, |
| { |
| "first": "Eva", |
| "middle": [], |
| "last": "Schlinger", |
| "suffix": "" |
| }, |
| { |
| "first": "Dan", |
| "middle": [], |
| "last": "Garrette", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1906.01502" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Telmo Pires, Eva Schlinger, and Dan Garrette. 2019. How multilingual is multilingual bert? arXiv preprint arXiv:1906.01502.", |
| "links": null |
| }, |
| "BIBREF32": { |
| "ref_id": "b32", |
| "title": "Lost word is lost world-a study of malayalam", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "Ri Prasanth", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Language in India", |
| "volume": "", |
| "issue": "6", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "RI Prasanth. 2016. Lost word is lost world-a study of malayalam. Language in India, 16(6).", |
| "links": null |
| }, |
| "BIBREF33": { |
| "ref_id": "b33", |
| "title": "Semantic cosine similarity", |
| "authors": [ |
| { |
| "first": "Faisal", |
| "middle": [], |
| "last": "Rahutomo", |
| "suffix": "" |
| }, |
| { |
| "first": "Teruaki", |
| "middle": [], |
| "last": "Kitasuka", |
| "suffix": "" |
| }, |
| { |
| "first": "Masayoshi", |
| "middle": [], |
| "last": "Aritsugi", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "The 7th International Student Conference on Advanced Science and Technology ICAST", |
| "volume": "4", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Faisal Rahutomo, Teruaki Kitasuka, and Masayoshi Aritsugi. 2012. Semantic cosine similarity. In The 7th International Student Conference on Advanced Science and Technology ICAST, volume 4.", |
| "links": null |
| }, |
| "BIBREF34": { |
| "ref_id": "b34", |
| "title": "Malayalam wordnet", |
| "authors": [ |
| { |
| "first": "S", |
| "middle": [], |
| "last": "Rajendran", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Soman", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "The WordNet in Indian Languages", |
| "volume": "", |
| "issue": "", |
| "pages": "119--145", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "S Rajendran and KP Soman. 2017. Malayalam wordnet. In The WordNet in Indian Languages, pages 119-145. Springer.", |
| "links": null |
| }, |
| "BIBREF35": { |
| "ref_id": "b35", |
| "title": "Sentencebert: Sentence embeddings using siamese bertnetworks", |
| "authors": [ |
| { |
| "first": "Nils", |
| "middle": [], |
| "last": "Reimers", |
| "suffix": "" |
| }, |
| { |
| "first": "Iryna", |
| "middle": [], |
| "last": "Gurevych", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1908.10084" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nils Reimers and Iryna Gurevych. 2019. Sentence- bert: Sentence embeddings using siamese bert- networks. arXiv preprint arXiv:1908.10084.", |
| "links": null |
| }, |
| "BIBREF36": { |
| "ref_id": "b36", |
| "title": "Malayalam spell checker", |
| "authors": [ |
| { |
| "first": "T", |
| "middle": [], |
| "last": "Santhosh", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Varghese", |
| "suffix": "" |
| }, |
| { |
| "first": "R", |
| "middle": [], |
| "last": "Sulochana", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Kumar", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "Proceedings of the International Conference on Universal Knowledge and Language", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "T Santhosh, KG Varghese, R Sulochana, and R Ku- mar. 2002. Malayalam spell checker. In Proceed- ings of the International Conference on Univer- sal Knowledge and Language-2002.", |
| "links": null |
| }, |
| "BIBREF37": { |
| "ref_id": "b37", |
| "title": "English to malayalam transliteration using sequence labeling approach", |
| "authors": [ |
| { |
| "first": "Sumaja", |
| "middle": [], |
| "last": "Sasidharan", |
| "suffix": "" |
| }, |
| { |
| "first": "K", |
| "middle": [ |
| "P" |
| ], |
| "last": "Loganathan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Soman", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "International Journal of Recent Trends in Engineering", |
| "volume": "1", |
| "issue": "2", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Sumaja Sasidharan, R Loganathan, and KP So- man. 2009. English to malayalam translitera- tion using sequence labeling approach. Interna- tional Journal of Recent Trends in Engineering, 1(2):170.", |
| "links": null |
| }, |
| "BIBREF38": { |
| "ref_id": "b38", |
| "title": "Deep learning based spell checker for malayalam language", |
| "authors": [ |
| { |
| "first": "S", |
| "middle": [], |
| "last": "Sooraj", |
| "suffix": "" |
| }, |
| { |
| "first": "Anand", |
| "middle": [], |
| "last": "Manjusha", |
| "suffix": "" |
| }, |
| { |
| "first": "K", |
| "middle": [ |
| "P" |
| ], |
| "last": "Kumar", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Soman", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "Journal of Intelligent & Fuzzy Systems", |
| "volume": "34", |
| "issue": "3", |
| "pages": "1427--1434", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "S Sooraj, K Manjusha, M Anand Kumar, and KP Soman. 2018. Deep learning based spell checker for malayalam language. Journal of In- telligent & Fuzzy Systems, 34(3):1427-1434.", |
| "links": null |
| }, |
| "BIBREF39": { |
| "ref_id": "b39", |
| "title": "A collaborative platform for multilingual ontology development", |
| "authors": [ |
| { |
| "first": "Ahmed", |
| "middle": [], |
| "last": "Tawfik", |
| "suffix": "" |
| }, |
| { |
| "first": "Fausto", |
| "middle": [], |
| "last": "Giunchiglia", |
| "suffix": "" |
| }, |
| { |
| "first": "Vincenzo", |
| "middle": [], |
| "last": "Maltese", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "World Academy of Science, Engineering and Technology", |
| "volume": "8", |
| "issue": "12", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ahmed Tawfik, Fausto Giunchiglia, and Vincenzo Maltese. 2014. A collaborative platform for mul- tilingual ontology development. World Academy of Science, Engineering and Technology, 8(12):1.", |
| "links": null |
| }, |
| "BIBREF40": { |
| "ref_id": "b40", |
| "title": "Quality assurance in translation. a process-oriented approach", |
| "authors": [ |
| { |
| "first": "Titela", |
| "middle": [], |
| "last": "V\u00eelceanu", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Romanian Journal of English Studies", |
| "volume": "14", |
| "issue": "1", |
| "pages": "141--146", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Titela V\u00eelceanu. 2017. Quality assurance in trans- lation. a process-oriented approach. Romanian Journal of English Studies, 14(1):141-146.", |
| "links": null |
| }, |
| "BIBREF41": { |
| "ref_id": "b41", |
| "title": "Reconceptualizing the digital divide", |
| "authors": [ |
| { |
| "first": "Mark", |
| "middle": [], |
| "last": "Warschauer", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "First monday", |
| "volume": "", |
| "issue": "7", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Mark Warschauer. 2002. Reconceptualizing the digital divide. First monday, 7(7).", |
| "links": null |
| }, |
| "BIBREF42": { |
| "ref_id": "b42", |
| "title": "Navigating the job search as a low-resourced job seeker", |
| "authors": [ |
| { |
| "first": "Earnest", |
| "middle": [], |
| "last": "Wheeler", |
| "suffix": "" |
| }, |
| { |
| "first": "Tawanna", |
| "middle": [ |
| "R" |
| ], |
| "last": "Dillahunt", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Earnest Wheeler and Tawanna R Dillahunt. 2018. Navigating the job search as a low-resourced job seeker. In Proceedings of the 2018 CHI Confer- ence on Human Factors in Computing Systems, page 48. ACM.", |
| "links": null |
| }, |
| "BIBREF43": { |
| "ref_id": "b43", |
| "title": "Learnware: on the future of machine learning", |
| "authors": [ |
| { |
| "first": "Zhi-Hua", |
| "middle": [], |
| "last": "Zhou", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Frontiers Comput. Sci", |
| "volume": "10", |
| "issue": "4", |
| "pages": "589--590", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Zhi-Hua Zhou. 2016. Learnware: on the future of machine learning. Frontiers Comput. Sci., 10(4):589-590.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF0": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "Language wise Internet users based on the study conducted by KPMG crowdsourcing application(Brabham, 2013)." |
| }, |
| "FIGREF1": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "Example for Empty example sentence Figure 3: Example for poorly defined gloss" |
| }, |
| "FIGREF2": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "Example for poorly defined pos Figure 5: Example for unknown characters in gloss" |
| }, |
| "FIGREF4": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "Examples for semantic errors" |
| }, |
| "FIGREF5": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "Summary of the similarity agreement of MWN between PWN and HWN" |
| }, |
| "FIGREF6": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "Distribution of similarity agreement within MWN Figure 10: Distribution of similarity of MWN between PWN and HWN" |
| }, |
| "TABREF2": { |
| "num": null, |
| "html": null, |
| "text": "", |
| "content": "<table><tr><td>:</td><td>Schema errors in Malayalam lexical-</td></tr><tr><td colspan=\"2\">semantic resource</td></tr></table>", |
| "type_str": "table" |
| } |
| } |
| } |
| } |
