Base_JSON/prefixS/json/sigtyp/2022.sigtyp-1.7.json

{
    "paper_id": "2022",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T15:21:02.112254Z"
    },
    "title": "The SIGTYP 2022 Shared Task on the Prediction of Cognate Reflexes",
    "authors": [
        {
            "first": "Johann-Mattis",
            "middle": [],
            "last": "List",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "Ekaterina",
            "middle": [],
            "last": "Vylomova",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "Robert",
            "middle": [],
            "last": "Forkel",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "Nathan",
            "middle": [
                "W"
            ],
            "last": "Hill",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "University of Dublin D ETH Z\u00fcrich",
                "location": {}
            },
            "email": ""
        },
        {
            "first": "Ryan",
            "middle": [
                "D"
            ],
            "last": "Cotterell",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "D W Mpi-Eva",
            "middle": [],
            "last": "Leipzig",
            "suffix": "",
            "affiliation": {},
            "email": ""
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "This study describes the structure and the results of the SIGTYP 2022 shared task on the prediction of cognate reflexes from multilingual wordlists. We asked participants to submit systems that would predict words in individual languages with the help of cognate words from related languages. Training and surprise data were based on standardized multilingual wordlists from several language families. Four teams submitted a total of eight systems, including both neural and non-neural systems, as well as systems adjusted to the task and systems using more general settings. While all systems showed a rather promising performance, reflecting the overwhelming regularity of sound change, the best performance throughout was achieved by a system based on convolutional networks originally designed for image restoration.",
    "pdf_parse": {
        "paper_id": "2022",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "This study describes the structure and the results of the SIGTYP 2022 shared task on the prediction of cognate reflexes from multilingual wordlists. We asked participants to submit systems that would predict words in individual languages with the help of cognate words from related languages. Training and surprise data were based on standardized multilingual wordlists from several language families. Four teams submitted a total of eight systems, including both neural and non-neural systems, as well as systems adjusted to the task and systems using more general settings. While all systems showed a rather promising performance, reflecting the overwhelming regularity of sound change, the best performance throughout was achieved by a system based on convolutional networks originally designed for image restoration.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "In historical-comparative linguistics, scholars typically assemble words from related languages into cognate sets. In contrast to the notion of cognacy in language teaching and synchronic NLP applications, cognate sets are understood as sets of words that share a common origin regardless of their meaning in historical-comparative linguistics and that should not contain borrowed words. The individual members of a cognate set are typically called cognate reflexes or simply reflexes (Trask, 2000, 278) . Cognate reflexes typically show regular sound correspondences. This means that one can define a mapping across the individual phoneme systems of the individual languages. Thus, English t typically corresponds to a German ts (compare ten vs. zehn), and English d corresponds to German t (compare dove vs. Taube). The mappings often depend on certain contextual conditions and may differ, depending on the position in which they occur in a word. With the help of regular sound correspondences, linguists can often predict fairly well how the cognate counterpart of a word in one language might sound in another language. However, prediction by linguists rarely takes only one language pair into account. The more reflexes a cognate set has in different languages, the easier it is to predict reflexes in individual languages.",
                "cite_spans": [
                    {
                        "start": 485,
                        "end": 503,
                        "text": "(Trask, 2000, 278)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "In its simplest form, the data we need for the task of reflex prediction is a table in which each column represents a different language and each row a different cognate set. We also assume that word forms (or \"reflexes\" of a cognate set) are represented in standardized phonetic transcriptions (such as the International Phonetic Alphabet). Whenever a reflex in a specific language is missing, this reflex can in theory be predicted with the help of the remaining reflexes. As an example, consider Table  2 , showing reflexes of cognate sets in German, English, and Dutch. Since the reflex for the BELLY cognate sets is missing in English, we could try and predict it from known correspondences to German and Dutch. The correct prediction would be bouk. This form has been still preserved for some time in English in the meaning of \"torso\", going back to Old English b\u016bk \"belly\" (Pfeifer, 1993) , although it has nowadays come out of use. When provided with more data of this kind, one can build a model that would be able to predict an English form given a German and a Dutch form, as well as a German form, given a Dutch and an English form, and so on. Note that not all cognate sets in real-life data will have reflexes for all words. Thus, we know about English bouk from dialect records, but without dialects or written sources from Middle English, we could only rely on prediction itself in order to guess how the word would sound if it would have been retained.",
                "cite_spans": [
                    {
                        "start": 880,
                        "end": 895,
                        "text": "(Pfeifer, 1993)",
                        "ref_id": "BIBREF56"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 499,
                        "end": 507,
                        "text": "Table  2",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "The Reflex Prediction Task",
                "sec_num": "1.1"
            },
            {
                "text": "Since predictions for words that have been completely lost cannot be evaluated directly, we will base our task on the prediction of artificially ex- * wangbai Wang and Wang (2004) v1.0 Sino-Tibetan 10 4356 658 Table 1 : Training and surprise data data used in our study. Datasets with identifiers preceded by an asterisk are those in which we automatically searched for cognates. The remaining datasets all provided expert cognates, which we used for the shared task. All datasets are archived with Zenodo, and the supplementary material provides a direct reference to their Zenodo DOI and their GitHub repository URLs.",
                "cite_spans": [
                    {
                        "start": 159,
                        "end": 179,
                        "text": "Wang and Wang (2004)",
                        "ref_id": "BIBREF63"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 210,
                        "end": 217,
                        "text": "Table 1",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "The Reflex Prediction Task",
                "sec_num": "1.1"
            },
            {
                "text": "cluded word forms. Thus, we first take a dataset with cognates in a few related languages, and then artificially delete some of the words in the datasets, using varying proportions. When training a model to predict the missing word forms, we can then compare the predicted words directly with the words we have deleted automatically (List, 2019a) . A special case of the reflex prediction task, supervised phonological reconstruction, focuses on the prediction of words in ancestral languages, thus mimicking the process of phonological reconstruction as one of the key aspects of the traditional comparative method (Weiss, 2015 ). While we predict reflexes in any language in the generic reflex prediction task, in automated phonological reconstruction we predict one specific reflex of a cognate set, viz. the form in the ancestral language. Apart from the restriction in scope, however, the two tasks do not differ much, and most methods which solve the one task could also be used to solve the other. List (2019a) in order to propose a new framework for supervised phonological reconstruction and cognate reflex prediction which they expand by enriching phonetic alignment analyses in such a way that contextual information can be taken account.",
                "cite_spans": [
                    {
                        "start": 333,
                        "end": 346,
                        "text": "(List, 2019a)",
                        "ref_id": "BIBREF39"
                    },
                    {
                        "start": 616,
                        "end": 628,
                        "text": "(Weiss, 2015",
                        "ref_id": "BIBREF64"
                    },
                    {
                        "start": 1005,
                        "end": 1017,
                        "text": "List (2019a)",
                        "ref_id": "BIBREF39"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Reflex Prediction Task",
                "sec_num": "1.1"
            },
            {
                "text": "Cognate Set German English Dutch ASH a S E aeS A s BITE b ai s @ n b ai t b Ei t @ BELLY b au x - b oei k",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Reflex Prediction Task",
                "sec_num": "1.1"
            },
            {
                "text": "For traditional as well as modern approaches to reflex prediction, there are a couple of challenges that algorithmic solutions need to account for. Since there is no one-to-one sound match between the sound in English and the sound in Dutch, the prediction has to be based on the conditioning context, which is notoriously difficult to handle in computational approaches.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Difficulties of Reflex Prediction",
                "sec_num": "1.3"
            },
            {
                "text": "A further difficulty consists in the sparsity and the patchiness of the data. Data are sparse with respect to the number of cognate sets which we can use to train computers or humans. Even for wellestablished language groups, etymological dictionaries, which list more than 1000 reconstructed items are quite rare. Apart from being sparse, data are also patchy. Only a very small amount of the proto-forms listed in etymological dictionaries is reflected in the majority of the branches, and an even smaller amount has survived without notable irregularities in the sound changes or the morphology of the word forms. Thus, even if one works with datasets consisting of large numbers of related words, there will always be situations in which important reflexes are missing and at times only one witness may be left that we can use for the prediction of the cognate reflex in question.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Difficulties of Reflex Prediction",
                "sec_num": "1.3"
            },
            {
                "text": "Data for the shared task were taken from the Lexibank repository, which offers wordlists from 100 standardized datasets (List et al. 2022a , https://github.com/lexibank/ lexibank-analysed). In this repository, a large collection of datasets with cognate sets provided by experts and phonetic transcriptions added by the Lexibank team are provided. An even larger number of datasets has only standardized phonetic transcriptions but no cognate judgments. Since cognate detection methods work well by now, we can determine the cognates specifically for shallower language families with quite some confidence; this enabled us to assemble a larger amount of datasets from different language families and either use cognate sets provided by experts or inferring cognates ourselves, using state-of-the-art methods for automated cognate detection implemented in the LingPy software library (List and Forkel, 2021) .",
                "cite_spans": [
                    {
                        "start": 120,
                        "end": 138,
                        "text": "(List et al. 2022a",
                        "ref_id": null
                    },
                    {
                        "start": 883,
                        "end": 906,
                        "text": "(List and Forkel, 2021)",
                        "ref_id": "BIBREF41"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Materials",
                "sec_num": "2.1"
            },
            {
                "text": "For each the training and the surprise phase, 10 datasets were selected. Following the Lexibank workflow for the curation of lexical wordlists, all datasets were curated on GitHub and additionally archived with Zenodo. Standardization of the data included mapping the language names to Glottolog (Hammarstr\u00f6m et al., 2021) , linking the concept elicitation glosses to the Concepticon reference catalog (https://concepticon.clld. org, List et al. 2022c), and adding standardized phonetic transcriptions, following the B(road)IPA system of the Cross-Linguistic Transcription Systems reference catalog (https://clts.clld. org, Anderson et al. 2018) , with the help of orthography profiles (Moran and Cysouw, 2018) . Since only a smaller number of the datasets came along with suitable cognate judgments needed for the cognate reflex prediction task, cognates were automatically inferred with standard settings, using a variant of the LexStat algorithm for automatic cognate detection (List, 2012a ) that searches for partial rather than full-word cognates (List et al., 2016) . Searching for partial cognates is justified, since both the identification of regular sound correspondences and the prediction of cognate reflexes can only be carried out on material that is entirely cognate (Schweikhard and List, 2020) . Since full-word cognates may often contain non-cognate material, the prediction of full cognates would unnecessarily exacerbate the reflex prediction task, adding a random component that cannot be handled algorithmically in a principled way. In all cases, we excluded all singleton cognate sets (cognate sets that occur only in one language), since these cannot be used in our prediction experiments. Table 1 lists all datasets for the test and training phase along with some basic statistics.",
                "cite_spans": [
                    {
                        "start": 296,
                        "end": 322,
                        "text": "(Hammarstr\u00f6m et al., 2021)",
                        "ref_id": null
                    },
                    {
                        "start": 624,
                        "end": 645,
                        "text": "Anderson et al. 2018)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 686,
                        "end": 710,
                        "text": "(Moran and Cysouw, 2018)",
                        "ref_id": "BIBREF52"
                    },
                    {
                        "start": 981,
                        "end": 993,
                        "text": "(List, 2012a",
                        "ref_id": "BIBREF36"
                    },
                    {
                        "start": 1053,
                        "end": 1072,
                        "text": "(List et al., 2016)",
                        "ref_id": "BIBREF46"
                    },
                    {
                        "start": 1283,
                        "end": 1311,
                        "text": "(Schweikhard and List, 2020)",
                        "ref_id": "BIBREF59"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 1715,
                        "end": 1728,
                        "text": "Table 1 lists",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Materials",
                "sec_num": "2.1"
            },
            {
                "text": "The datasets were used as the basis for the data used for test and training during our shared task. For this purpose, each dataset was split into five training and test partitions in which the data retained for testing was varied, starting from a proportion of 10% retained for testing (proportion 0.1), followed by 20% (proportion 0.2), 30% (proportion 0.3), 40% (proportion 0.4), and finally 50% (proportion 0.5). The training data was not modified further and used as primary input for the training phase of all systems. The test data, however, was artificially constructed from the test partition. We first iterated over all cognate sets and then created individual test sets from each cognate set iterating over all words in a cognate set and deleting each word in a row. For a cognate set of n words, this would result in n test cases, in which each word in each language would have to be predicted one time.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Materials",
                "sec_num": "2.1"
            },
            {
                "text": "Among the most commonly used evaluation measures for the word prediction task is the edit distance, which computes the number of operations needed in order to convert the predicted word into the attested word (Levenshtein, 1965) . In its primary form, the edit distance is an integer. In order to normalize the measure, correcting for a bias resulting from the length of the compared strings, scholars have proposed to divide the distance by the length of the largest string (Holman et al., 2008) , or by the mean length of both strings being compared (Nerbonne et al., 1999) . A further possibility closer to notions of distance in bioinformatics, which we used in our shared task, is to divide the edit distance by the length of the alignment of both strings. The normalized edit distance then corresponds to the normalized Hamming distance between two aligned sequences (Hamming, 1950) , or -when subtracting from 1 -to the notion of percentage identity in evolutionary biology (Raghava and Barton, 2006) . It is, however, important to note that actual differences in these normalization procedures are usually small.",
                "cite_spans": [
                    {
                        "start": 209,
                        "end": 228,
                        "text": "(Levenshtein, 1965)",
                        "ref_id": "BIBREF34"
                    },
                    {
                        "start": 475,
                        "end": 496,
                        "text": "(Holman et al., 2008)",
                        "ref_id": "BIBREF26"
                    },
                    {
                        "start": 552,
                        "end": 575,
                        "text": "(Nerbonne et al., 1999)",
                        "ref_id": "BIBREF53"
                    },
                    {
                        "start": 873,
                        "end": 888,
                        "text": "(Hamming, 1950)",
                        "ref_id": "BIBREF23"
                    },
                    {
                        "start": 981,
                        "end": 1007,
                        "text": "(Raghava and Barton, 2006)",
                        "ref_id": "BIBREF57"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Evaluation",
                "sec_num": "2.2.1"
            },
            {
                "text": "The edit distance, both normalized and unnormalized, has been employed in many word prediction and phonological reconstruction experiments as the basic evaluation measure for the prediction accuracy (Meloni et al., 2021; Bouchard-C\u00f4t\u00e9 et al., 2013) . Its clearest shortcoming lies in the fact that it only accounts for surface differences between prediction and attested words (also called 'phenotypic differences' by Lass 1997) , while structural aspects (called 'genotypic differences' by Lass 1997) are ignored. Thus, if a method mistakenly maps a certain sound x to a certain sound y in all cases in which the x occurs, the edit distance will treat each occurrence of the error independently and may therefore provide drastically lowered results. It would, therefore, be good to account for the relative regularity of the co-occurrence of x and y. List (2019b) proposes to compute B-Cubed F-scores (Amig\u00f3 et al., 2009) from the aligned predicted and attested words. B-cubed F-scores only check for the regularity of occurrences. This results in scores of 1 (indicating complete identity) for sequence pairs like abbc compared with 1223. Indeed, both sequences are structurally completely identical since a simple mapping between the symbols in both sequences can convert one string into the other and vice versa. If a method has systematic errors but otherwise does a good job in prediction, B-Cubed F-Scores penalize results less strongly than edit distance. As a final evaluation score, we followed Fourrier et al. (2021) in providing BLEU scores (Papineni et al., 2002) . These scores are usually used to investigate how well an automated translation corresponds to the translated target test. BLEU scores and B-Cubed F-Scores range from 0 to 1, with 1 indicating perfect agreement, the normalized edit distance ranges between 1 (maximal difference) and 0 (string identity).",
                "cite_spans": [
                    {
                        "start": 199,
                        "end": 220,
                        "text": "(Meloni et al., 2021;",
                        "ref_id": "BIBREF51"
                    },
                    {
                        "start": 221,
                        "end": 248,
                        "text": "Bouchard-C\u00f4t\u00e9 et al., 2013)",
                        "ref_id": "BIBREF8"
                    },
                    {
                        "start": 418,
                        "end": 428,
                        "text": "Lass 1997)",
                        "ref_id": "BIBREF33"
                    },
                    {
                        "start": 491,
                        "end": 501,
                        "text": "Lass 1997)",
                        "ref_id": "BIBREF33"
                    },
                    {
                        "start": 902,
                        "end": 922,
                        "text": "(Amig\u00f3 et al., 2009)",
                        "ref_id": "BIBREF2"
                    },
                    {
                        "start": 1505,
                        "end": 1527,
                        "text": "Fourrier et al. (2021)",
                        "ref_id": "BIBREF18"
                    },
                    {
                        "start": 1553,
                        "end": 1576,
                        "text": "(Papineni et al., 2002)",
                        "ref_id": "BIBREF54"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Evaluation",
                "sec_num": "2.2.1"
            },
            {
                "text": "Our baselines were taken from the reflex prediction framework by List et al. (2022b) . This framework consists of four major stages. In stage (1), cognate sets are aligned with the help of standard methods for multiple phonetic alignment analyses (List, 2012b) . In stage (2), alignments are trimmed by merging all columns in the alignment in which the attested languages all show a gap with their preceding column. As a result, a word like Latin cen\u0101re [k e: n a: r E] would be rendered as [k e: n a: r.E], when being aligned with Spanish cenar [T e n a R], since the final [E] in Latin corresponds to a gap in Spanish and could therefore not be predicted (see List et al. 2022b for details on this procedure). In stage (3), alignments are enriched by coding for potentially conditioning context, which is added to the alignments in the form of additional rows. In stage (4), the individual alignment columns are converted to a matrix from which a classifier can be trained. During prediction, cognate sets fed to the algorithm are again being aligned and enriched, but the trimming procedure is not needed, since it only relates to the target language that one wants to predict. Based on this general framework, we created two baselines, one primary baseline that uses the correspondence pattern recognition (CORPAR) method by List (2019a) as a classifier, and one extended baseline which predicts words with the help of a support vector machine (SVM, see List et al. 2022b for details on both systems). From previous studies on supervised phonological reconstruction we know that the SVM variant of the framework outperforms the CORPAR classifier clearly, although differences are not extremely high (ibid.).",
                "cite_spans": [
                    {
                        "start": 65,
                        "end": 84,
                        "text": "List et al. (2022b)",
                        "ref_id": "BIBREF44"
                    },
                    {
                        "start": 247,
                        "end": 260,
                        "text": "(List, 2012b)",
                        "ref_id": "BIBREF37"
                    },
                    {
                        "start": 575,
                        "end": 578,
                        "text": "[E]",
                        "ref_id": null
                    },
                    {
                        "start": 662,
                        "end": 679,
                        "text": "List et al. 2022b",
                        "ref_id": "BIBREF44"
                    },
                    {
                        "start": 1458,
                        "end": 1475,
                        "text": "List et al. 2022b",
                        "ref_id": "BIBREF44"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Baselines",
                "sec_num": "2.2.2"
            },
            {
                "text": "We created specific software package that allows to (1) automatically download the data in the particular versions of the individual CLDF datasets which we used, (2) create the test and training data, (3) apply the baseline methods to the data, and (4) carry out the evaluation. The software package is written in Python and can be accessed both using the commandline and from within Python scripts. It is curated on GitHub and archived with Zenodo (see Section Supplementary Material for details). Different versions were created in order to first release the training data (version 1.1), followed by the release of the surprise data (version 1.2), and finally followed by the release of the official results of the evaluation (version 1.4, providing extended evaluations in contrast to the version 1.3 planned earlier).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Implementation",
                "sec_num": "2.2.3"
            },
            {
                "text": "Major dependencies of the software package are LingPy (List and Forkel, 2021) , used for the com-putation of the edit distance and of phonetic alignments, Lingrex (List and Forkel, 2022) , providing access to the baseline method for cognate reflex prediction, Scikit-learn (Pedregosa et al., 2011) , providing access to support vector machines, and Matplotlib (Hunter, 2007) , used for plotting.",
                "cite_spans": [
                    {
                        "start": 54,
                        "end": 77,
                        "text": "(List and Forkel, 2021)",
                        "ref_id": "BIBREF41"
                    },
                    {
                        "start": 163,
                        "end": 186,
                        "text": "(List and Forkel, 2022)",
                        "ref_id": "BIBREF42"
                    },
                    {
                        "start": 273,
                        "end": 297,
                        "text": "(Pedregosa et al., 2011)",
                        "ref_id": "BIBREF55"
                    },
                    {
                        "start": 360,
                        "end": 374,
                        "text": "(Hunter, 2007)",
                        "ref_id": "BIBREF27"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Implementation",
                "sec_num": "2.2.3"
            },
            {
                "text": "Four teams submitted their systems for our shared task. Since these systems are described in individual papers (Kirov et al., 2022; J\u00e4ger, 2022; Tresoldi, 2022 ; Celano, 2022), we will only briefly present their main features here. Team CrossLingference, represented by Gerhard J\u00e4ger (University T\u00fcbingen), provided a workflow J\u00e4ger 2022, implemented in the JULIA programming language, that makes specifically use of Bayesian phylogenetic inference. In contrast to the remaining systems submitted to our shared task, J\u00e4ger's approach takes phylogenetic information into account, extending an earlier workflow for phonological reconstruction (J\u00e4ger, 2019) . Team Mockingbird, represented by Christo Kirov, Richard Sproat, and Alexander Gutkin (Google Research), provided two models for the prediction of cognate reflexes. The first model, the NEIGHBOR TRANSFORMER MODEL, was originally designed to find problems in the readings of Japanese place names spelled in kanji (Jones et al., 2022) , and is based on the popular transformer architecture (Vaswani et al., 2017) , which was specifically adjusted for the task. Since the training data would be too small for the transformer model, the authors augmented it with new instances generated by randomly sampling subsets of a corresponding cognate set. In addition to that, they also enriched each set with synthetic instances using n-gram language modelling. The second model, the IMAGE INPAINTING MODEL, compares the cognate reflex prediction task to the task of restoring corrupted parts of a 2D image, in which dimensions correspond to languages and cognate phonemic representations. The restoration is achieved with the help of convolutional neural networks (Liu et al., 2018) . For this model, no data augmentation steps were undertaken. The authors provide four model configurations of the neighbor model, with the first three (N1-A, N1-B, and N1-C) differing in the number of training steps and not being publicly released, while the last one (N2), which was only applied to the 0.1 proportion of the data, being pub-licly released. For the image inpainting model, one configuration was provided (I1). Team Leipzig, represented by Giuseppe G. A. Celano (University Leipzig), provided a TRANS-FORMER-based architecture with character and position embeddings for the prediction of cognate reflexes (Vaswani et al., 2017) , in which language information was one-hot encoded and the model was trained on individual reflex pairs on each language independently. In order to predict a word from several reflexes in different languages, the system first predicts individual target tensors of probabilities for each attested reflex and then averages them to produce the prediction. Team CEoT, represented by Tiago Tresoldi (Uppsala University), provided a workflow that predicts cognate reflexes based on phonetic alignments (Tresoldi, 2022) , which is quite similar to the extended baselines of our shared task (List et al., 2022b) . In contrast to our baseline approaches, their system EXTALIGN-RF skips the trimming procedure (stage 2), varies the techniques for alignment enrichment by taking preceding and following context into account (stage 3), and uses a random forests classifier rather than a support vector machine.",
                "cite_spans": [
                    {
                        "start": 111,
                        "end": 131,
                        "text": "(Kirov et al., 2022;",
                        "ref_id": "BIBREF32"
                    },
                    {
                        "start": 132,
                        "end": 144,
                        "text": "J\u00e4ger, 2022;",
                        "ref_id": "BIBREF30"
                    },
                    {
                        "start": 145,
                        "end": 159,
                        "text": "Tresoldi, 2022",
                        "ref_id": "BIBREF61"
                    },
                    {
                        "start": 641,
                        "end": 654,
                        "text": "(J\u00e4ger, 2019)",
                        "ref_id": "BIBREF29"
                    },
                    {
                        "start": 968,
                        "end": 988,
                        "text": "(Jones et al., 2022)",
                        "ref_id": "BIBREF28"
                    },
                    {
                        "start": 1044,
                        "end": 1066,
                        "text": "(Vaswani et al., 2017)",
                        "ref_id": "BIBREF62"
                    },
                    {
                        "start": 1710,
                        "end": 1728,
                        "text": "(Liu et al., 2018)",
                        "ref_id": "BIBREF48"
                    },
                    {
                        "start": 2351,
                        "end": 2373,
                        "text": "(Vaswani et al., 2017)",
                        "ref_id": "BIBREF62"
                    },
                    {
                        "start": 2871,
                        "end": 2887,
                        "text": "(Tresoldi, 2022)",
                        "ref_id": "BIBREF61"
                    },
                    {
                        "start": 2958,
                        "end": 2978,
                        "text": "(List et al., 2022b)",
                        "ref_id": "BIBREF44"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Systems",
                "sec_num": "3"
            },
            {
                "text": "While all teams tried hard to provide results for all of their systems, some results could not be computed in time, to be included in the shared task. All teams were asked to share their data in such a way that users can easily replicate the results and also apply their methods to new data. Unfortunately, there was no time for the team organizing the shared task to individually check all systems with respect to replicability and transparency. The team checked, however, that all systems were properly archived with repositories offering long-term storage of data, such as Zenodo, and we communicated the importance of replication with all authors.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Systems",
                "sec_num": "3"
            },
            {
                "text": "Given that we measure system performance with four evaluation measures (edit distance, normalized edit distance, B-Cubed F-Scores, and BLEU scores adjusted for word prediction), one might expect that systems perform differently with respect to different evaluation measures. As can be seen from the results in Table 3 , however, the results are rather clearly favoring the system I1 by the team Mockingbird as the winner in almost all proportions. The only case where the Mockingbird I1 ing (proportion 0.5), where the JULIA system by the CrossLingference team shows the best performance with respect to the B-Cubed F-Scores. Since B-Cubed F-Scores emphasize the systematicity of the prediction quality rather than the accuracy in individual cases, we can see that the JULIA system copes better with systematic aspects of the word prediction tasks in those cases, where the data for the training of the system is limited. That the different scoring systems show at least some degree of independence can also be seen in Figure 2 , which shows results for the 10% partition, where the JU-LIA system performs worst with respect to edit distances and BLEU scores, while showing a better performance than N2, TRANSFORMER, and the baseline in B-Cubed F-Scores. While the SVM baseline shows a surprisingly good performance on the lowest proportion of data excluded and retained for testing (proportion 0.1), it looses ground with more data excluded for testing. Here, the N1-A and N1-B systems, again from the Mockingbird team, show the best performance. Table 4 provides the aggregated ranks for the normalized edit distance, the B-Cubed F-Scores, and the BLEU scores for all systems obtained for all splits of the data. The classical edit distance was excluded in this overview, since it correlates highly with the normalized edit distance and would therefore artificially increase the overall ranks of systems performing well in this regard. Furthermore, the N2 system by the Mockingbird team was excluded in this analysis, since results could only be provided for the smallest proportion of words retained for testing (proportion 0.1). For each of the five splits of the data and for each of the methods, we ranked the systems according to their performance and later calculated the average of all ranks for each system on each of the three evaluation methods. The aggregated ranks, in which all three evaluation measures are ranked equally, allow us to rank the overall performance of all systems. It shows the overall superiority of the I1 system of the Mockingbird team, followed by the teams' N1-A and N1-B methods. The SVM baseline and the N1-C method by team Mockingbird follow on places four and five. At the end of these ranks are the EXTALIGN-RF system by team CeOT, the JU-LIA system by Team CrossLingference, followed by the simple baseline and the TRANSFORMER approach of team Leipzig.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 310,
                        "end": 317,
                        "text": "Table 3",
                        "ref_id": "TABREF4"
                    },
                    {
                        "start": 1019,
                        "end": 1027,
                        "text": "Figure 2",
                        "ref_id": null
                    },
                    {
                        "start": 1548,
                        "end": 1555,
                        "text": "Table 4",
                        "ref_id": "TABREF5"
                    }
                ],
                "eq_spans": [],
                "section": "Results",
                "sec_num": "4"
            },
            {
                "text": "Overall, all systems do quite a good job at recovering unknown words from their cognate sets, specifically in those cases, where only a small part of the test data was retained for the evaluation process. Judging from our practical experience and independently published results on word prediction experiments (List et al., 2022b; Bodt and List, 2022) , B-Cubed F-Scores higher than 0.7 and average edit distances of about 1 provide a good starting point for computer-assisted approaches and can already provide active help in various practical annotation tasks in historical linguistics. Thus, scholars working on the reconstruction of certain language families could use predicted proto-forms and later manually correct them, or field workers could use automatically predicted words when trying to elicit specific lexical items to search for cognate words that might have shifted their meanings.",
                "cite_spans": [
                    {
                        "start": 310,
                        "end": 330,
                        "text": "(List et al., 2022b;",
                        "ref_id": "BIBREF44"
                    },
                    {
                        "start": 331,
                        "end": 351,
                        "text": "Bodt and List, 2022)",
                        "ref_id": "BIBREF7"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Results",
                "sec_num": "4"
            },
            {
                "text": "It was one of the crucial insights made by historical linguists in the early 19th century (Grimm, 1822; Rask, 1818) , that sound change proceeds in a surprisingly regular, systematic manner, affecting all sounds in the lexicon of a language that recur in similar phonotactic positions. Without the systematicity and regularity of sound change, it would not be possible to predict the pronunciation of words in one language based on the pronunciation of cognate words in related languages. While it has been known for a long time to linguists that these kinds of predictions can be made on the basis of historical language comparison, the task of cognate reflex prediction has only recently attracted the attention of scholars working in the field of Natural Language Processing and computational linguistics.",
                "cite_spans": [
                    {
                        "start": 90,
                        "end": 103,
                        "text": "(Grimm, 1822;",
                        "ref_id": null
                    },
                    {
                        "start": 104,
                        "end": 115,
                        "text": "Rask, 1818)",
                        "ref_id": "BIBREF58"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussion",
                "sec_num": "5"
            },
            {
                "text": "With our shared task on cognate reflex prediction, we hoped to achieve two major goals. On the one hand, we wanted to highlight the importance of classical scholarship for computational applications in historical linguistics and linguistic Figure 2 : Results for the surprise dataset of the 0.1 proportion, with 10% of data retained for testing.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 240,
                        "end": 248,
                        "text": "Figure 2",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Discussion",
                "sec_num": "5"
            },
            {
                "text": "typology, showing that quite a few problems which are up to today exclusively solved manually might profit from computational treatment. On the other hand, we wanted to trigger the interest of scholars with diverse backgrounds in this task, assembling teams that address the problem with different strategies that might inspire each other and help to lead to largely improved methods in the future.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussion",
                "sec_num": "5"
            },
            {
                "text": "With the four teams that participated, we have seen an interesting and diverse assembly of systems that all deal with the cognate reflex prediction task. While two teams made use of state-of-the-art machine learning methods based on neural networks (team Mockingbird and team Leipzig), two teams represented systems based on workflows using classical approaches in the emerging discipline of computational historical linguistics (team CrossLingference and team CEoT), using phonetic alignments, and -in the case of team CrossLingference -even Bayesian methods for phylogenetic reconstruction. From the overall performance of the systems in our shared task, we can see that some of the neural approaches outperform the more targeted solutions. Given differences in the performance with respect to the evaluation methods, which highlight different aspects of prediction accuracy, however, we could also see that targeted methods like the Julia method by CrossLingference or the extended Baseline come very close to the best neural systems, and even outperform them at times.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussion",
                "sec_num": "5"
            }
        ],
        "back_matter": [
            {
                "text": "This study was partially supported by the Max Planck Society Research Grant \"Beyond CALC: Computer-Assisted Approaches to Human Prehistory, Linguistic Typology, and Human Cognition (CALC\u00b3)\", awarded to Johann-Mattis List ( -2024. We thank Cl\u00e9mentine Fourrier for comments and suggestions on the evaluation procedure, and we thank all participants for their feedback during the task.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Acknowledgements",
                "sec_num": null
            },
            {
                "text": "Data and code for the shared task along with results for all systems are curated GitHub (https:// github.com/sigtyp/ST2022, Version 1.4) and have been archived with Zenodo (https:// doi.org/10.5281/zenodo.6586772).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Supplementary Material",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "A sociolinguistic research among selected groups in Western Arunachal Pradesh highlighting Monpa",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Binny",
                        "suffix": ""
                    }
                ],
                "year": 2005,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Binny et al Abraham. 2005. A sociolinguistic research among selected groups in Western Arunachal Pradesh highlighting Monpa. Unpublished manuscript.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Bai Dialect Survey. SIL International",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Bryan Allen",
                        "suffix": ""
                    }
                ],
                "year": 2007,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Bryan Allen. 2007. Bai Dialect Survey. SIL Interna- tional, Dallas.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "A comparison of extrinsic clustering evaluation metrics based on formal constraints",
                "authors": [
                    {
                        "first": "Enrique",
                        "middle": [],
                        "last": "Amig\u00f3",
                        "suffix": ""
                    },
                    {
                        "first": "Julio",
                        "middle": [],
                        "last": "Gonzalo",
                        "suffix": ""
                    },
                    {
                        "first": "Javier",
                        "middle": [],
                        "last": "Artiles",
                        "suffix": ""
                    },
                    {
                        "first": "Felisa",
                        "middle": [],
                        "last": "Verdejo",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "Information Retrieval",
                "volume": "12",
                "issue": "4",
                "pages": "461--486",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Enrique Amig\u00f3, Julio Gonzalo, Javier Artiles, and Fe- lisa Verdejo. 2009. A comparison of extrinsic clus- tering evaluation metrics based on formal constraints. Information Retrieval, 12(4):461-486.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "A Cross-Linguistic Database of Phonetic Transcription Systems",
                "authors": [
                    {
                        "first": "Cormac",
                        "middle": [],
                        "last": "Anderson",
                        "suffix": ""
                    },
                    {
                        "first": "Tiago",
                        "middle": [],
                        "last": "Tresoldi",
                        "suffix": ""
                    },
                    {
                        "first": "Thiago",
                        "middle": [],
                        "last": "Costa Chacon",
                        "suffix": ""
                    },
                    {
                        "first": "Anne-Maria",
                        "middle": [],
                        "last": "Fehn",
                        "suffix": ""
                    },
                    {
                        "first": "Mary",
                        "middle": [],
                        "last": "Walworth",
                        "suffix": ""
                    },
                    {
                        "first": "Robert",
                        "middle": [],
                        "last": "Forkel",
                        "suffix": ""
                    },
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2018,
                "venue": "Yearbook of the Pozna\u0144 Linguistic Meeting",
                "volume": "4",
                "issue": "",
                "pages": "21--53",
                "other_ids": {
                    "DOI": [
                        "10.2478/yplm-2018-0002"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Cormac Anderson, Tiago Tresoldi, Thiago Costa Cha- con, Anne-Maria Fehn, Mary Walworth, Robert Forkel, and Johann-Mattis List. 2018. A Cross- Linguistic Database of Phonetic Transcription Sys- tems. Yearbook of the Pozna\u0144 Linguistic Meeting, 4(1):21-53.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "of Sociolinguistic Survey of Northern Pakistan",
                "authors": [
                    {
                        "first": "C",
                        "middle": [],
                        "last": "Peter",
                        "suffix": ""
                    },
                    {
                        "first": "Carla",
                        "middle": [
                            "F"
                        ],
                        "last": "Backstrom",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Radloff",
                        "suffix": ""
                    }
                ],
                "year": 1992,
                "venue": "",
                "volume": "2",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Peter C. Backstrom and Carla F. Radloff. 1992. Lan- guages of Northern Areas, volume 2 of Sociolinguis- tic Survey of Northern Pakistan. National Institute of Pakistan Studies, Islamabad.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "Cognate production using character-based machine translation",
                "authors": [
                    {
                        "first": "Lisa",
                        "middle": [],
                        "last": "Beinborn",
                        "suffix": ""
                    },
                    {
                        "first": "Torsten",
                        "middle": [],
                        "last": "Zesch",
                        "suffix": ""
                    },
                    {
                        "first": "Iryna",
                        "middle": [],
                        "last": "Gurevych",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Proceedings of the Sixth International Joint Conference on Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "883--891",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Lisa Beinborn, Torsten Zesch, and Iryna Gurevych. 2013. Cognate production using character-based machine translation. In Proceedings of the Sixth International Joint Conference on Natural Language Processing, pages 883-891.",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "A Combined Comparative and Phylogenetic Analysis of the Chapacuran Language Family",
                "authors": [
                    {
                        "first": "Joshua",
                        "middle": [],
                        "last": "Birchall",
                        "suffix": ""
                    },
                    {
                        "first": "Michael",
                        "middle": [],
                        "last": "Dunn",
                        "suffix": ""
                    },
                    {
                        "first": "Simon",
                        "middle": [
                            "J"
                        ],
                        "last": "Greenhill",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "International Journal of American Linguistics",
                "volume": "82",
                "issue": "3",
                "pages": "255--284",
                "other_ids": {
                    "DOI": [
                        "10.1086/687383"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Joshua Birchall, Michael Dunn, and Simon J. Greenhill. 2016. A Combined Comparative and Phylogenetic Analysis of the Chapacuran Language Family. Inter- national Journal of American Linguistics, 82(3):255- 284.",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Reflex prediction. A case study of Western Kho-Bwa",
                "authors": [
                    {
                        "first": "Timotheus",
                        "middle": [
                            "Adrianus"
                        ],
                        "last": "Bodt",
                        "suffix": ""
                    },
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2022,
                "venue": "Diachronica",
                "volume": "39",
                "issue": "1",
                "pages": "1--38",
                "other_ids": {
                    "DOI": [
                        "10.1075/dia.20009.bod"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Timotheus Adrianus Bodt and Johann-Mattis List. 2022. Reflex prediction. A case study of Western Kho-Bwa. Diachronica, 39(1):1-38.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "Automated reconstruction of ancient languages using probabilistic models of sound change",
                "authors": [
                    {
                        "first": "Alexandre",
                        "middle": [],
                        "last": "Bouchard-C\u00f4t\u00e9",
                        "suffix": ""
                    },
                    {
                        "first": "David",
                        "middle": [],
                        "last": "Hall",
                        "suffix": ""
                    },
                    {
                        "first": "Thomas",
                        "middle": [
                            "L"
                        ],
                        "last": "Griffiths",
                        "suffix": ""
                    },
                    {
                        "first": "Dan",
                        "middle": [],
                        "last": "Klein",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Proceedings of the National Academy of Sciences of the United States of America",
                "volume": "110",
                "issue": "11",
                "pages": "4224--4229",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Alexandre Bouchard-C\u00f4t\u00e9, David Hall, Thomas L. Grif- fiths, and Dan Klein. 2013. Automated reconstruc- tion of ancient languages using probabilistic mod- els of sound change. Proceedings of the National Academy of Sciences of the United States of America, 110(11):4224-4229.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "A sociolinguistic survey of six Berta speech varieties in Ethiopia. SIL International",
                "authors": [
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Nate",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Bremer",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Nate D. Bremer. 2016. A sociolinguistic survey of six Berta speech varieties in Ethiopia. SIL International, Addis Ababa.",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "H\u00e0ny\u01d4 f\u0101ngy\u012bn z\u00echu\u00ec \u6c49\u8bed\u65b9\u97f3\u5b57\u6c47",
                "authors": [],
                "year": 1962,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Beijing University B\u011bij\u012bng D\u00e0xu\u00e9. 1962. H\u00e0ny\u01d4 f\u0101ngy\u012bn z\u00echu\u00ec \u6c49\u8bed\u65b9\u97f3\u5b57\u6c47 [Chinese dialect character pro- nunciation list].",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "Sui dialect research. SIL International",
                "authors": [],
                "year": 2015,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Andy Castro and Xingwen Pan, editors. 2015. Sui di- alect research. SIL International, Guizhou.",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "A Transformer architecture for the prediction of cognate reflexes",
                "authors": [
                    {
                        "first": "G",
                        "middle": [
                            "A"
                        ],
                        "last": "Giuseppe",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Celano",
                        "suffix": ""
                    }
                ],
                "year": 2022,
                "venue": "The Fourth Workshop on Computational Typology and Multilingual NLP, Online. Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Giuseppe G. A. Celano. 2022. A Transformer architec- ture for the prediction of cognate reflexes. In The Fourth Workshop on Computational Typology and Multilingual NLP, Online. Association for Computa- tional Linguistics.",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "Proto-uto-aztecans on their way to the proto-aztecan homeland: linguistic evidence",
                "authors": [
                    {
                        "first": "Albert",
                        "middle": [],
                        "last": "Davletshin",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "Journal of Language Relationship",
                "volume": "1",
                "issue": "8",
                "pages": "75--92",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Albert Davletshin. 2012. Proto-uto-aztecans on their way to the proto-aztecan homeland: linguistic evi- dence. Journal of Language Relationship, 1(8):75- 92.",
                "links": null
            },
            "BIBREF15": {
                "ref_id": "b15",
                "title": "A lexical comparison of the Palaung dialects spoken in China",
                "authors": [
                    {
                        "first": "Sujaritlak",
                        "middle": [],
                        "last": "Deepadung",
                        "suffix": ""
                    },
                    {
                        "first": "Supakit",
                        "middle": [],
                        "last": "Buakaw",
                        "suffix": ""
                    },
                    {
                        "first": "Ampika",
                        "middle": [],
                        "last": "Rattanapitak",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "Mon-Khmer Studies",
                "volume": "44",
                "issue": "",
                "pages": "19--38",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Sujaritlak Deepadung, Supakit Buakaw, and Ampika Rattanapitak. 2015. A lexical comparison of the Palaung dialects spoken in China, Myanmar, and Thailand. Mon-Khmer Studies, 44:19-38.",
                "links": null
            },
            "BIBREF16": {
                "ref_id": "b16",
                "title": "Word prediction in computational historical linguistics",
                "authors": [
                    {
                        "first": "Peter",
                        "middle": [],
                        "last": "Dekker",
                        "suffix": ""
                    },
                    {
                        "first": "Willem",
                        "middle": [],
                        "last": "Zuidema",
                        "suffix": ""
                    }
                ],
                "year": 2021,
                "venue": "Journal of Language Modelling",
                "volume": "8",
                "issue": "2",
                "pages": "295--336",
                "other_ids": {
                    "DOI": [
                        "10.15398/jlm.v8i2.268"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Peter Dekker and Willem Zuidema. 2021. Word predic- tion in computational historical linguistics. Journal of Language Modelling, 8(2):295-336.",
                "links": null
            },
            "BIBREF17": {
                "ref_id": "b17",
                "title": "Ethiosemitic languages: Classifications and classification determinants. Ampersand",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Tekabe Legesse Feleke",
                        "suffix": ""
                    }
                ],
                "year": 2021,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {
                    "DOI": [
                        "10.1016/j.amper.2021.100074"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Tekabe Legesse Feleke. 2021. Ethiosemitic languages: Classifications and classification determinants. Am- persand, page 100074.",
                "links": null
            },
            "BIBREF18": {
                "ref_id": "b18",
                "title": "Can cognate prediction be modelled as a lowresource machine translation task?",
                "authors": [
                    {
                        "first": "Cl\u00e9mentine",
                        "middle": [],
                        "last": "Fourrier",
                        "suffix": ""
                    },
                    {
                        "first": "Rachel",
                        "middle": [],
                        "last": "Bawden",
                        "suffix": ""
                    },
                    {
                        "first": "Beno\u00eet",
                        "middle": [],
                        "last": "Sagot",
                        "suffix": ""
                    }
                ],
                "year": 2021,
                "venue": "Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021",
                "volume": "",
                "issue": "",
                "pages": "847--861",
                "other_ids": {
                    "DOI": [
                        "10.18653/v1/2021.findings-acl.75"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Cl\u00e9mentine Fourrier, Rachel Bawden, and Beno\u00eet Sagot. 2021. Can cognate prediction be modelled as a low- resource machine translation task? In Findings of the Association for Computational Linguistics: ACL- IJCNLP 2021, pages 847-861, Online. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF19": {
                "ref_id": "b19",
                "title": "Materials for an Etymological Dictionary of Burmish",
                "authors": [
                    {
                        "first": "Xun",
                        "middle": [],
                        "last": "Gong",
                        "suffix": ""
                    },
                    {
                        "first": "Nathan",
                        "middle": [],
                        "last": "Hill",
                        "suffix": ""
                    }
                ],
                "year": 2020,
                "venue": "Zenodo",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {
                    "DOI": [
                        "10.5281/zenodo.4311182"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Xun Gong and Nathan Hill. 2020. Materials for an Ety- mological Dictionary of Burmish. Zenodo, Geneva.",
                "links": null
            },
            "BIBREF20": {
                "ref_id": "b20",
                "title": "Bantu Basic Vocabulary Database",
                "authors": [
                    {
                        "first": "J",
                        "middle": [],
                        "last": "Simon",
                        "suffix": ""
                    },
                    {
                        "first": "Russell D",
                        "middle": [],
                        "last": "Greenhill",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Gray",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Simon J Greenhill and Russell D Gray. 2015. Bantu Basic Vocabulary Database.",
                "links": null
            },
            "BIBREF23": {
                "ref_id": "b23",
                "title": "Error detection and error detection codes",
                "authors": [
                    {
                        "first": "Richard",
                        "middle": [
                            "W"
                        ],
                        "last": "Hamming",
                        "suffix": ""
                    }
                ],
                "year": 1950,
                "venue": "Bell System Technical Journal",
                "volume": "29",
                "issue": "2",
                "pages": "147--160",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Richard W. Hamming. 1950. Error detection and er- ror detection codes. Bell System Technical Journal, 29(2):147-160.",
                "links": null
            },
            "BIBREF24": {
                "ref_id": "b24",
                "title": "Bangime: Secret language, language isolate, or language island?",
                "authors": [
                    {
                        "first": "Abbie",
                        "middle": [],
                        "last": "Hantgan",
                        "suffix": ""
                    },
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2018,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Abbie Hantgan and Johann-Mattis List. 2018. Bangime: Secret language, language isolate, or language is- land?",
                "links": null
            },
            "BIBREF25": {
                "ref_id": "b25",
                "title": "Japanese dialects",
                "authors": [
                    {
                        "first": "Shir\u014d",
                        "middle": [],
                        "last": "Hattori",
                        "suffix": ""
                    }
                ],
                "year": 1973,
                "venue": "Diachronic, areal and typological linguistics, number 11 in Current Trends in Linguistics",
                "volume": "",
                "issue": "",
                "pages": "368--400",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Shir\u014d Hattori. 1973. Japanese dialects. In Henry M. Hoenigswald and Robert H. Langacre, editors, Di- achronic, areal and typological linguistics, num- ber 11 in Current Trends in Linguistics, pages 368- 400. Mouton, The Hague and Paris.",
                "links": null
            },
            "BIBREF26": {
                "ref_id": "b26",
                "title": "Advances in automated language classification",
                "authors": [
                    {
                        "first": "Eric",
                        "middle": [
                            "W"
                        ],
                        "last": "Holman",
                        "suffix": ""
                    },
                    {
                        "first": "S\u00f8ren",
                        "middle": [],
                        "last": "Wichmann",
                        "suffix": ""
                    },
                    {
                        "first": "Cecil",
                        "middle": [
                            "H"
                        ],
                        "last": "Brown",
                        "suffix": ""
                    },
                    {
                        "first": "Viveka",
                        "middle": [],
                        "last": "Velupillai",
                        "suffix": ""
                    },
                    {
                        "first": "Andr\u00e9",
                        "middle": [],
                        "last": "M\u00fcller",
                        "suffix": ""
                    },
                    {
                        "first": "Dik",
                        "middle": [],
                        "last": "Bakker",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "Quantitative Investigations in Theoretical Linguistics",
                "volume": "",
                "issue": "",
                "pages": "40--43",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Eric W. Holman, S\u00f8ren Wichmann, Cecil H. Brown, Viveka Velupillai, Andr\u00e9 M\u00fcller, and Dik Bakker. 2008. Advances in automated language classification. In Antti Arppe, Kaius Sinnem\u00e4ki, and Urpu Nikann, editors, Quantitative Investigations in Theoretical Linguistics, pages 40-43. University of Helsinki, Helsinki.",
                "links": null
            },
            "BIBREF27": {
                "ref_id": "b27",
                "title": "Matplotlib: A 2D graphics environment",
                "authors": [
                    {
                        "first": "D",
                        "middle": [],
                        "last": "John",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Hunter",
                        "suffix": ""
                    }
                ],
                "year": 2007,
                "venue": "Computing In Science & Engineering",
                "volume": "9",
                "issue": "3",
                "pages": "90--95",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "John D. Hunter. 2007. Matplotlib: A 2D graphics en- vironment. Computing In Science & Engineering, 9(3):90-95.",
                "links": null
            },
            "BIBREF28": {
                "ref_id": "b28",
                "title": "Helpful neighbors: Leveraging geographic neighbors to aid in placename pronunciation",
                "authors": [
                    {
                        "first": "Llion",
                        "middle": [],
                        "last": "Jones",
                        "suffix": ""
                    },
                    {
                        "first": "Richard",
                        "middle": [],
                        "last": "Sproat",
                        "suffix": ""
                    },
                    {
                        "first": "Haruko",
                        "middle": [],
                        "last": "Ishikawa",
                        "suffix": ""
                    }
                ],
                "year": 2022,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Llion Jones, Richard Sproat, and Haruko Ishikawa. 2022. Helpful neighbors: Leveraging geographic neighbors to aid in placename pronunciation. In preparation.",
                "links": null
            },
            "BIBREF29": {
                "ref_id": "b29",
                "title": "Computational historical linguistics",
                "authors": [
                    {
                        "first": "Gerhard",
                        "middle": [],
                        "last": "J\u00e4ger",
                        "suffix": ""
                    }
                ],
                "year": 2019,
                "venue": "Theoretical Linguistics",
                "volume": "45",
                "issue": "3-4",
                "pages": "151--182",
                "other_ids": {
                    "DOI": [
                        "10.1515/tl-2019-0011"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Gerhard J\u00e4ger. 2019. Computational historical linguis- tics. Theoretical Linguistics, 45(3-4):151-182.",
                "links": null
            },
            "BIBREF30": {
                "ref_id": "b30",
                "title": "Bayesian phylogenetic cognate prediction",
                "authors": [
                    {
                        "first": "Gerhard",
                        "middle": [],
                        "last": "J\u00e4ger",
                        "suffix": ""
                    }
                ],
                "year": 2022,
                "venue": "The Fourth Workshop on Computational Typology and Multilingual NLP, Online. Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Gerhard J\u00e4ger. 2022. Bayesian phylogenetic cognate prediction. In The Fourth Workshop on Computa- tional Typology and Multilingual NLP, Online. Asso- ciation for Computational Linguistics.",
                "links": null
            },
            "BIBREF31": {
                "ref_id": "b31",
                "title": "The significance of word lists",
                "authors": [
                    {
                        "first": "Brett",
                        "middle": [],
                        "last": "Kessler",
                        "suffix": ""
                    }
                ],
                "year": 2001,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Brett Kessler. 2001. The significance of word lists. CSLI Publications, Stanford.",
                "links": null
            },
            "BIBREF32": {
                "ref_id": "b32",
                "title": "Mockingbird at the SIGTYP 2022 Shared Task: Two types of models for the prediction of cognate reflexes",
                "authors": [
                    {
                        "first": "Christo",
                        "middle": [],
                        "last": "Kirov",
                        "suffix": ""
                    },
                    {
                        "first": "Richard",
                        "middle": [],
                        "last": "Sproat",
                        "suffix": ""
                    },
                    {
                        "first": "Alexander",
                        "middle": [],
                        "last": "Gutkin",
                        "suffix": ""
                    }
                ],
                "year": 2022,
                "venue": "The Fourth Workshop on Computational Typology and Multilingual NLP",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Christo Kirov, Richard Sproat, and Alexander Gutkin. 2022. Mockingbird at the SIGTYP 2022 Shared Task: Two types of models for the prediction of cognate reflexes. In The Fourth Workshop on Computational Typology and Multilingual NLP, Online. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF33": {
                "ref_id": "b33",
                "title": "Historical linguistics and language change",
                "authors": [
                    {
                        "first": "Roger",
                        "middle": [],
                        "last": "Lass",
                        "suffix": ""
                    }
                ],
                "year": 1997,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Roger Lass. 1997. Historical linguistics and language change. Cambridge University Press, Cambridge.",
                "links": null
            },
            "BIBREF34": {
                "ref_id": "b34",
                "title": "Dvoi\u010dnye kody s ispravleniem vypadenij, vstavok i zame\u0161\u010denij simvolov",
                "authors": [
                    {
                        "first": ".",
                        "middle": [
                            "I"
                        ],
                        "last": "Vladimir",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Levenshtein",
                        "suffix": ""
                    }
                ],
                "year": 1965,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Vladimir. I. Levenshtein. 1965. Dvoi\u010dnye kody s is- pravleniem vypadenij, vstavok i zame\u0161\u010denij simvolov [binary codes with correction of deletions, insertions and replacements].",
                "links": null
            },
            "BIBREF36": {
                "ref_id": "b36",
                "title": "Automatic detection of cognates in multilingual wordlists",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "Proceedings of the EACL 2012 Joint Workshop of Visualization of Linguistic Patterns and Uncovering Language History from Multilingual Resources",
                "volume": "",
                "issue": "",
                "pages": "117--125",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List. 2012a. LexStat. Automatic detec- tion of cognates in multilingual wordlists. In Pro- ceedings of the EACL 2012 Joint Workshop of Visu- alization of Linguistic Patterns and Uncovering Lan- guage History from Multilingual Resources, pages 117-125, Stroudsburg.",
                "links": null
            },
            "BIBREF37": {
                "ref_id": "b37",
                "title": "SCA: Phonetic alignment based on sound classes",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "New directions in logic, language, and computation",
                "volume": "",
                "issue": "",
                "pages": "32--51",
                "other_ids": {
                    "DOI": [
                        "10.1007/978-3-642-31467-4_3"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List. 2012b. SCA: Phonetic alignment based on sound classes. In Marija Slavkovik and Dan Lassiter, editors, New directions in logic, language, and computation, pages 32-51. Springer, Berlin and Heidelberg.",
                "links": null
            },
            "BIBREF38": {
                "ref_id": "b38",
                "title": "Investigating the impact of sample size on cognate detection",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Journal of Language Relationship",
                "volume": "11",
                "issue": "",
                "pages": "91--101",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List. 2014. Investigating the impact of sample size on cognate detection. Journal of Lan- guage Relationship, 11:91-101.",
                "links": null
            },
            "BIBREF39": {
                "ref_id": "b39",
                "title": "Automatic inference of sound correspondence patterns across multiple languages",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2019,
                "venue": "Computational Linguistics",
                "volume": "45",
                "issue": "1",
                "pages": "137--161",
                "other_ids": {
                    "DOI": [
                        "10.1162/coli_a_00344"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List. 2019a. Automatic inference of sound correspondence patterns across multiple lan- guages. Computational Linguistics, 45(1):137-161.",
                "links": null
            },
            "BIBREF40": {
                "ref_id": "b40",
                "title": "Beyond Edit Distances: Comparing linguistic reconstruction systems",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2019,
                "venue": "Theoretical Linguistics",
                "volume": "45",
                "issue": "3-4",
                "pages": "1--10",
                "other_ids": {
                    "DOI": [
                        "10.1515/tl-2019-0016"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List. 2019b. Beyond Edit Distances: Comparing linguistic reconstruction systems. Theo- retical Linguistics, 45(3-4):1-10.",
                "links": null
            },
            "BIBREF41": {
                "ref_id": "b41",
                "title": "LingPy. A Python library for quantitative tasks in historical linguistics",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    },
                    {
                        "first": "Robert",
                        "middle": [],
                        "last": "Forkel",
                        "suffix": ""
                    }
                ],
                "year": 2021,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List and Robert Forkel. 2021. LingPy. A Python library for quantitative tasks in histor- ical linguistics [Software Library, Version 2.6.9].",
                "links": null
            },
            "BIBREF42": {
                "ref_id": "b42",
                "title": "LingRex: Linguistic reconstruction with LingPy",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    },
                    {
                        "first": "Robert",
                        "middle": [],
                        "last": "Forkel",
                        "suffix": ""
                    }
                ],
                "year": 2022,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List and Robert Forkel. 2022. LingRex: Linguistic reconstruction with LingPy [Software Li- brary, Version 1.2].",
                "links": null
            },
            "BIBREF43": {
                "ref_id": "b43",
                "title": "Johannes Englisch, and Russell D. Gray. 2022a. Lexibank, A public repository of standardized wordlists with computed phonological and lexical features",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    },
                    {
                        "first": "Robert",
                        "middle": [],
                        "last": "Forkel",
                        "suffix": ""
                    },
                    {
                        "first": "Simon",
                        "middle": [
                            "J"
                        ],
                        "last": "Greenhill",
                        "suffix": ""
                    },
                    {
                        "first": "Christoph",
                        "middle": [],
                        "last": "Rzymski",
                        "suffix": ""
                    }
                ],
                "year": null,
                "venue": "Scientific Data",
                "volume": "",
                "issue": "",
                "pages": "1--31",
                "other_ids": {
                    "DOI": [
                        "10.21203/rs.3.rs-870835/v1"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List, Robert Forkel, Simon J. Greenhill, Christoph Rzymski, Johannes Englisch, and Rus- sell D. Gray. 2022a. Lexibank, A public repository of standardized wordlists with computed phonological and lexical features. Scientific Data, pages 1-31.",
                "links": null
            },
            "BIBREF44": {
                "ref_id": "b44",
                "title": "A new framework for fast automated phonological reconstruction using trimmed alignments and sound correspondence patterns",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    },
                    {
                        "first": "Nathan",
                        "middle": [
                            "W"
                        ],
                        "last": "Hill",
                        "suffix": ""
                    },
                    {
                        "first": "Robert",
                        "middle": [],
                        "last": "Forkel",
                        "suffix": ""
                    }
                ],
                "year": 2022,
                "venue": "Proceedings of the 3rd Workshop on Computational Approaches to Historical Language Change",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List, Nathan W. Hill, and Robert Forkel. 2022b. A new framework for fast automated phono- logical reconstruction using trimmed alignments and sound correspondence patterns. In Proceedings of the 3rd Workshop on Computational Approaches to Historical Language Change, Dublin [Online].",
                "links": null
            },
            "BIBREF45": {
                "ref_id": "b45",
                "title": "Association for Computational Linguistics",
                "authors": [],
                "year": null,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Asso- ciation for Computational Linguistics.",
                "links": null
            },
            "BIBREF46": {
                "ref_id": "b46",
                "title": "Using sequence similarity networks to identify partial cognates in multilingual wordlists",
                "authors": [
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    },
                    {
                        "first": "Philippe",
                        "middle": [],
                        "last": "Lopez",
                        "suffix": ""
                    },
                    {
                        "first": "Eric",
                        "middle": [],
                        "last": "Bapteste",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "Proceedings of the Association of Computational Linguistics",
                "volume": "2",
                "issue": "",
                "pages": "599--605",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Johann-Mattis List, Philippe Lopez, and Eric Bapteste. 2016. Using sequence similarity networks to iden- tify partial cognates in multilingual wordlists. In Proceedings of the Association of Computational Linguistics 2016 (Volume 2: Short Papers), pages 599-605, Berlin. Association of Computational Lin- guistics.",
                "links": null
            },
            "BIBREF48": {
                "ref_id": "b48",
                "title": "Image inpainting for irregular holes using partial convolutions",
                "authors": [
                    {
                        "first": "Guilin",
                        "middle": [],
                        "last": "Liu",
                        "suffix": ""
                    },
                    {
                        "first": "Fitsum",
                        "middle": [
                            "A"
                        ],
                        "last": "Reda",
                        "suffix": ""
                    },
                    {
                        "first": "Kevin",
                        "middle": [
                            "J"
                        ],
                        "last": "Shih",
                        "suffix": ""
                    },
                    {
                        "first": "Ting-Chun",
                        "middle": [],
                        "last": "Wang",
                        "suffix": ""
                    },
                    {
                        "first": "Andrew",
                        "middle": [],
                        "last": "Tao",
                        "suffix": ""
                    },
                    {
                        "first": "Bryan",
                        "middle": [],
                        "last": "Catanzaro",
                        "suffix": ""
                    }
                ],
                "year": 2018,
                "venue": "Proceedings of the 15th European Conference on Computer Vision (ECCV 2018)",
                "volume": "",
                "issue": "",
                "pages": "89--105",
                "other_ids": {
                    "DOI": [
                        "10.1007/978-3-030-01252-6_6"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Guilin Liu, Fitsum A. Reda, Kevin J. Shih, Ting-Chun Wang, Andrew Tao, and Bryan Catanzaro. 2018. Im- age inpainting for irregular holes using partial convo- lutions. In Proceedings of the 15th European Con- ference on Computer Vision (ECCV 2018), pages 89-105, Munich, Germany. Springer International Publishing. Preprint.",
                "links": null
            },
            "BIBREF49": {
                "ref_id": "b49",
                "title": "A view on Proto-Karen phonology and lexicon",
                "authors": [
                    {
                        "first": "Theraphan",
                        "middle": [],
                        "last": "Luangthongkum",
                        "suffix": ""
                    }
                ],
                "year": 2019,
                "venue": "Journal of the Southeast Asian Linguistics Society",
                "volume": "12",
                "issue": "1",
                "pages": "i--lii",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Theraphan Luangthongkum. 2019. A view on Proto- Karen phonology and lexicon. Journal of the South- east Asian Linguistics Society, 12(1):i-lii.",
                "links": null
            },
            "BIBREF50": {
                "ref_id": "b50",
                "title": "A phonological reconstruction of Proto Northern Burmic. Phd, The University of Texas",
                "authors": [
                    {
                        "first": "Noel",
                        "middle": [],
                        "last": "Walter Mann",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Noel Walter Mann. 1998. A phonological reconstruc- tion of Proto Northern Burmic. Phd, The University of Texas, Arlington.",
                "links": null
            },
            "BIBREF51": {
                "ref_id": "b51",
                "title": "Ab antiquo: Neural proto-language reconstruction",
                "authors": [
                    {
                        "first": "Carlo",
                        "middle": [],
                        "last": "Meloni",
                        "suffix": ""
                    },
                    {
                        "first": "Shauli",
                        "middle": [],
                        "last": "Ravfogel",
                        "suffix": ""
                    },
                    {
                        "first": "Yoav",
                        "middle": [],
                        "last": "Goldberg",
                        "suffix": ""
                    }
                ],
                "year": 2021,
                "venue": "Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
                "volume": "",
                "issue": "",
                "pages": "4460--4473",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Carlo Meloni, Shauli Ravfogel, and Yoav Goldberg. 2021. Ab antiquo: Neural proto-language recon- struction. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Tech- nologies, pages 4460-4473, Online. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF52": {
                "ref_id": "b52",
                "title": "The Unicode Cookbook for Linguists: Managing writing systems using orthography profiles",
                "authors": [
                    {
                        "first": "Steven",
                        "middle": [],
                        "last": "Moran",
                        "suffix": ""
                    },
                    {
                        "first": "Michael",
                        "middle": [],
                        "last": "Cysouw",
                        "suffix": ""
                    }
                ],
                "year": 2018,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Steven Moran and Michael Cysouw. 2018. The Unicode Cookbook for Linguists: Managing writing systems using orthography profiles. Language Science Press, Berlin.",
                "links": null
            },
            "BIBREF53": {
                "ref_id": "b53",
                "title": "Edit distance and dialect proximity",
                "authors": [
                    {
                        "first": "John",
                        "middle": [],
                        "last": "Nerbonne",
                        "suffix": ""
                    },
                    {
                        "first": "Wilbert",
                        "middle": [],
                        "last": "Heeringa",
                        "suffix": ""
                    },
                    {
                        "first": "Peter",
                        "middle": [],
                        "last": "Kleiweg",
                        "suffix": ""
                    }
                ],
                "year": 1999,
                "venue": "Time warps, string edits, and macromolecules. The theory and practice of sequence comparison",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "John Nerbonne, Wilbert Heeringa, and Peter Kleiweg. 1999. Edit distance and dialect proximity. In David Sankoff and Joseph. B. Kruskal, editors, Time warps, string edits, and macromolecules. The theory and practice of sequence comparison, reprint edition, pages V-XV. CSLI Publications, Stanford.",
                "links": null
            },
            "BIBREF54": {
                "ref_id": "b54",
                "title": "Bleu: a method for automatic evaluation of machine translation",
                "authors": [
                    {
                        "first": "Kishore",
                        "middle": [],
                        "last": "Papineni",
                        "suffix": ""
                    },
                    {
                        "first": "Salim",
                        "middle": [],
                        "last": "Roukos",
                        "suffix": ""
                    },
                    {
                        "first": "Todd",
                        "middle": [],
                        "last": "Ward",
                        "suffix": ""
                    },
                    {
                        "first": "Wei-Jing",
                        "middle": [],
                        "last": "Zhu",
                        "suffix": ""
                    }
                ],
                "year": 2002,
                "venue": "Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "311--318",
                "other_ids": {
                    "DOI": [
                        "10.3115/1073083.1073135"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Kishore Papineni, Salim Roukos, Todd Ward, and Wei- Jing Zhu. 2002. Bleu: a method for automatic evalu- ation of machine translation. In Proceedings of the 40th Annual Meeting of the Association for Compu- tational Linguistics, pages 311-318, Philadelphia, Pennsylvania, USA. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF55": {
                "ref_id": "b55",
                "title": "Scikit-learn: Machine learning in Python",
                "authors": [
                    {
                        "first": "F",
                        "middle": [],
                        "last": "Pedregosa",
                        "suffix": ""
                    },
                    {
                        "first": "G",
                        "middle": [],
                        "last": "Varoquaux",
                        "suffix": ""
                    },
                    {
                        "first": "A",
                        "middle": [],
                        "last": "Gramfort",
                        "suffix": ""
                    },
                    {
                        "first": "V",
                        "middle": [],
                        "last": "Michel",
                        "suffix": ""
                    },
                    {
                        "first": "B",
                        "middle": [],
                        "last": "Thirion",
                        "suffix": ""
                    },
                    {
                        "first": "O",
                        "middle": [],
                        "last": "Grisel",
                        "suffix": ""
                    },
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Blondel",
                        "suffix": ""
                    },
                    {
                        "first": "P",
                        "middle": [],
                        "last": "Prettenhofer",
                        "suffix": ""
                    },
                    {
                        "first": "R",
                        "middle": [],
                        "last": "Weiss",
                        "suffix": ""
                    },
                    {
                        "first": "V",
                        "middle": [],
                        "last": "Dubourg",
                        "suffix": ""
                    },
                    {
                        "first": "J",
                        "middle": [],
                        "last": "Vanderplas",
                        "suffix": ""
                    },
                    {
                        "first": "A",
                        "middle": [],
                        "last": "Passos",
                        "suffix": ""
                    },
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Cournapeau",
                        "suffix": ""
                    },
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Brucher",
                        "suffix": ""
                    },
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Perrot",
                        "suffix": ""
                    },
                    {
                        "first": "E",
                        "middle": [],
                        "last": "Duchesnay",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "Journal of Machine Learning Research",
                "volume": "12",
                "issue": "",
                "pages": "2825--2830",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and E. Duch- esnay. 2011. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12:2825-2830.",
                "links": null
            },
            "BIBREF56": {
                "ref_id": "b56",
                "title": "Etymologisches W\u00f6rterbuch des Deutschen",
                "authors": [
                    {
                        "first": "Wolfgang",
                        "middle": [],
                        "last": "Pfeifer",
                        "suffix": ""
                    }
                ],
                "year": 1993,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Wolfgang Pfeifer. 1993. Etymologisches W\u00f6rterbuch des Deutschen, 2 edition. Akademie, Berlin.",
                "links": null
            },
            "BIBREF57": {
                "ref_id": "b57",
                "title": "Quantification of the variation in percentage identity for protein sequence alignments",
                "authors": [
                    {
                        "first": "G",
                        "middle": [
                            "P S"
                        ],
                        "last": "Raghava",
                        "suffix": ""
                    },
                    {
                        "first": "Geoffrey",
                        "middle": [
                            "J"
                        ],
                        "last": "Barton",
                        "suffix": ""
                    }
                ],
                "year": 2006,
                "venue": "BMC Bioinformatics",
                "volume": "",
                "issue": "415",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "G. P. S. Raghava and Geoffrey J. Barton. 2006. Quan- tification of the variation in percentage identity for protein sequence alignments. BMC Bioinformatics, 7(415).",
                "links": null
            },
            "BIBREF58": {
                "ref_id": "b58",
                "title": "Unders\u00f6gelse om det gamle Nordiske eller Islandske sprogs oprindelse. Gyldendalske Boghandlings Forlag",
                "authors": [
                    {
                        "first": "K",
                        "middle": [],
                        "last": "Rasmus",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Rask",
                        "suffix": ""
                    }
                ],
                "year": null,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Rasmus K. Rask. 1818. Unders\u00f6gelse om det gamle Nordiske eller Islandske sprogs oprindelse. Gylden- dalske Boghandlings Forlag, Copenhagen.",
                "links": null
            },
            "BIBREF59": {
                "ref_id": "b59",
                "title": "Developing an annotation framework for word formation processes in comparative linguistics",
                "authors": [
                    {
                        "first": "Nathanael",
                        "middle": [
                            "E"
                        ],
                        "last": "Schweikhard",
                        "suffix": ""
                    },
                    {
                        "first": "Johann-Mattis",
                        "middle": [],
                        "last": "List",
                        "suffix": ""
                    }
                ],
                "year": 2020,
                "venue": "SKASE Journal of Theoretical Linguistics",
                "volume": "17",
                "issue": "1",
                "pages": "2--26",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Nathanael E. Schweikhard and Johann-Mattis List. 2020. Developing an annotation framework for word for- mation processes in comparative linguistics. SKASE Journal of Theoretical Linguistics, 17(1):2-26.",
                "links": null
            },
            "BIBREF60": {
                "ref_id": "b60",
                "title": "The dictionary of historical and comparative linguistics",
                "authors": [
                    {
                        "first": "L",
                        "middle": [],
                        "last": "Robert",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Trask",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Robert L. Trask. 2000. The dictionary of historical and comparative linguistics. Edinburgh University Press, Edinburgh.",
                "links": null
            },
            "BIBREF61": {
                "ref_id": "b61",
                "title": "Approaching reflex predictions as a classification problem from extended phonological alignments",
                "authors": [
                    {
                        "first": "Tiago",
                        "middle": [],
                        "last": "Tresoldi",
                        "suffix": ""
                    }
                ],
                "year": 2022,
                "venue": "The Fourth Workshop on Computational Typology and Multilingual NLP, Online",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Tiago Tresoldi. 2022. Approaching reflex predictions as a classification problem from extended phonolog- ical alignments. In The Fourth Workshop on Com- putational Typology and Multilingual NLP, Online. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF62": {
                "ref_id": "b62",
                "title": "Attention is all you need",
                "authors": [
                    {
                        "first": "Ashish",
                        "middle": [],
                        "last": "Vaswani",
                        "suffix": ""
                    },
                    {
                        "first": "Noam",
                        "middle": [],
                        "last": "Shazeer",
                        "suffix": ""
                    },
                    {
                        "first": "Niki",
                        "middle": [],
                        "last": "Parmar",
                        "suffix": ""
                    },
                    {
                        "first": "Jakob",
                        "middle": [],
                        "last": "Uszkoreit",
                        "suffix": ""
                    },
                    {
                        "first": "Llion",
                        "middle": [],
                        "last": "Jones",
                        "suffix": ""
                    },
                    {
                        "first": "Aidan",
                        "middle": [
                            "N"
                        ],
                        "last": "Gomez",
                        "suffix": ""
                    },
                    {
                        "first": "\u0141ukasz",
                        "middle": [],
                        "last": "Kaiser",
                        "suffix": ""
                    },
                    {
                        "first": "Illia",
                        "middle": [],
                        "last": "Polosukhin",
                        "suffix": ""
                    }
                ],
                "year": 2017,
                "venue": "Proceedings of the 31st Conference on Neural Information Processing Systems (NIPS 2017)",
                "volume": "",
                "issue": "",
                "pages": "5998--6008",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, \u0141ukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Proceedings of the 31st Conference on Neural Information Processing Systems (NIPS 2017), pages 5998-6008, Long Beach, CA. Curran Associates Inc.",
                "links": null
            },
            "BIBREF63": {
                "ref_id": "b63",
                "title": "Basic words and language evolution",
                "authors": [
                    {
                        "first": "Feng",
                        "middle": [],
                        "last": "Wang",
                        "suffix": ""
                    },
                    {
                        "first": "William S.-Y.",
                        "middle": [],
                        "last": "Wang",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "Language and Linguistics",
                "volume": "5",
                "issue": "3",
                "pages": "643--662",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Feng Wang and William S.-Y. Wang. 2004. Basic words and language evolution. Language and Linguistics, 5(3):643-662.",
                "links": null
            },
            "BIBREF64": {
                "ref_id": "b64",
                "title": "The comparative method",
                "authors": [
                    {
                        "first": "Michael",
                        "middle": [],
                        "last": "Weiss",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "The Routledge handbook of historical linguistics",
                "volume": "",
                "issue": "",
                "pages": "127--145",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Michael Weiss. 2015. The comparative method. In Claire Bowern and Nicholas Evans, editors, The Rout- ledge handbook of historical linguistics, pages 127- 145. Routledge, New York.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF1": {
                "num": null,
                "text": "Major steps of the reflex prediction framework underlying the baseline.",
                "uris": null,
                "type_str": "figure"
            },
            "TABREF1": {
                "num": null,
                "type_str": "table",
                "text": "",
                "html": null,
                "content": "<table><tr><td>: Exemplary cognate reflexes in German, En-glish, and Dutch.</td></tr></table>"
            },
            "TABREF4": {
                "num": null,
                "type_str": "table",
                "text": "",
                "html": null,
                "content": "<table><tr><td>: Results for the varying proportions and our four evaluation measures, edit distance (ED), normalized edit distance (NED), B-Cubed F-scores (B-Cubes) and BLEU Scores (BLEU) on the surprise data. Cells shaded in gray highlight the best score obtained for a given proportion, bold font marks the second best score.</td></tr><tr><td>system does not show the best performance is the</td></tr><tr><td>test with 50% of the words being retained for test-</td></tr></table>"
            },
            "TABREF5": {
                "num": null,
                "type_str": "table",
                "text": "Overview of the average ranks of all nine systems for the different dataset proportions along with aggregated ranks.",
                "html": null,
                "content": "<table/>"
            }
        }
    }
}