Cross-linguistic annotation transfer in geoparsing experiments with Classical texts. Soffiantini, L. DH Benelux Journal, 6:155–168, 2024. ![Cross-linguistic annotation transfer in geoparsing experiments with Classical texts [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper abstract bibtex The Natural History is an encyclopedic work written by the Latin author Pliny the Elder (first century CE). In this extensive text in 37 books, geography plays a pivotal role, with hundreds of mentions of ancient place names. In this paper, a geoparsing experiment is conducted on the Natural History with the scope of automatically identifying and extracting place entities. To achieve this, we take advantage of state-of-the-art NLP models to develop a multistage pipeline involving English Named Entity Recognition, English-Latin sentence alignment, and entity projection. The paper demonstrates how cross-lingual annotation transfer can be applied from a translation in a modern language back to the original text in the context of low-/medium-resource languages, such as Latin. The efficacy of the proposed pipeline is evaluated through the use of both standard metrics and a comprehensive manual error analysis. Additionally, the results are compared to those obtained by other Latin NER tools. Both analyses demonstrate that the proposed methodology achieves a superior f1-score. Finally, the majority of place entities were automatically associated with unique identifiers that enable geolocation by the projection of pre-disambiguated annotations derived from another geo-spatial project.
@article{soffiantini_cross-linguistic_2024,
title = {Cross-linguistic annotation transfer in geoparsing experiments with {Classical} texts},
volume = {6},
url = {https://journal.dhbenelux.org/wp-content/uploads/2024/11/9_Soffiantini_individual.pdf},
abstract = {The Natural History is an encyclopedic work written by the Latin author Pliny the Elder (first century CE). In this extensive text in 37 books, geography plays a pivotal role, with hundreds of mentions of ancient place names. In this paper, a geoparsing experiment is conducted on the Natural History with the scope of automatically identifying and extracting place entities. To achieve this, we take advantage of state-of-the-art NLP models to develop a multistage pipeline involving English Named Entity Recognition, English-Latin sentence alignment, and entity projection. The paper demonstrates how cross-lingual annotation transfer can be applied from a translation in a modern language back to the original text in the context of low-/medium-resource languages, such as Latin. The efficacy of the proposed pipeline is evaluated through the use of both standard metrics and a comprehensive manual error analysis. Additionally, the results are compared to those obtained by other Latin NER tools. Both analyses demonstrate that the proposed methodology achieves a superior f1-score. Finally, the majority of place entities were automatically associated with unique identifiers that enable geolocation by the projection of pre-disambiguated annotations derived from another geo-spatial project.},
urldate = {2025-01-26},
journal = {DH Benelux Journal},
author = {Soffiantini, Laura},
year = {2024},
pages = {155--168},
}
Downloads: 0
{"_id":"6DxXCaDBFjngtj23d","bibbaseid":"soffiantini-crosslinguisticannotationtransferingeoparsingexperimentswithclassicaltexts-2024","author_short":["Soffiantini, L."],"bibdata":{"bibtype":"article","type":"article","title":"Cross-linguistic annotation transfer in geoparsing experiments with Classical texts","volume":"6","url":"https://journal.dhbenelux.org/wp-content/uploads/2024/11/9_Soffiantini_individual.pdf","abstract":"The Natural History is an encyclopedic work written by the Latin author Pliny the Elder (first century CE). In this extensive text in 37 books, geography plays a pivotal role, with hundreds of mentions of ancient place names. In this paper, a geoparsing experiment is conducted on the Natural History with the scope of automatically identifying and extracting place entities. To achieve this, we take advantage of state-of-the-art NLP models to develop a multistage pipeline involving English Named Entity Recognition, English-Latin sentence alignment, and entity projection. The paper demonstrates how cross-lingual annotation transfer can be applied from a translation in a modern language back to the original text in the context of low-/medium-resource languages, such as Latin. The efficacy of the proposed pipeline is evaluated through the use of both standard metrics and a comprehensive manual error analysis. Additionally, the results are compared to those obtained by other Latin NER tools. Both analyses demonstrate that the proposed methodology achieves a superior f1-score. Finally, the majority of place entities were automatically associated with unique identifiers that enable geolocation by the projection of pre-disambiguated annotations derived from another geo-spatial project.","urldate":"2025-01-26","journal":"DH Benelux Journal","author":[{"propositions":[],"lastnames":["Soffiantini"],"firstnames":["Laura"],"suffixes":[]}],"year":"2024","pages":"155–168","bibtex":"@article{soffiantini_cross-linguistic_2024,\n\ttitle = {Cross-linguistic annotation transfer in geoparsing experiments with {Classical} texts},\n\tvolume = {6},\n\turl = {https://journal.dhbenelux.org/wp-content/uploads/2024/11/9_Soffiantini_individual.pdf},\n\tabstract = {The Natural History is an encyclopedic work written by the Latin author Pliny the Elder (first century CE). In this extensive text in 37 books, geography plays a pivotal role, with hundreds of mentions of ancient place names. In this paper, a geoparsing experiment is conducted on the Natural History with the scope of automatically identifying and extracting place entities. To achieve this, we take advantage of state-of-the-art NLP models to develop a multistage pipeline involving English Named Entity Recognition, English-Latin sentence alignment, and entity projection. The paper demonstrates how cross-lingual annotation transfer can be applied from a translation in a modern language back to the original text in the context of low-/medium-resource languages, such as Latin. The efficacy of the proposed pipeline is evaluated through the use of both standard metrics and a comprehensive manual error analysis. Additionally, the results are compared to those obtained by other Latin NER tools. Both analyses demonstrate that the proposed methodology achieves a superior f1-score. Finally, the majority of place entities were automatically associated with unique identifiers that enable geolocation by the projection of pre-disambiguated annotations derived from another geo-spatial project.},\n\turldate = {2025-01-26},\n\tjournal = {DH Benelux Journal},\n\tauthor = {Soffiantini, Laura},\n\tyear = {2024},\n\tpages = {155--168},\n}\n\n\n\n","author_short":["Soffiantini, L."],"key":"soffiantini_cross-linguistic_2024","id":"soffiantini_cross-linguistic_2024","bibbaseid":"soffiantini-crosslinguisticannotationtransferingeoparsingexperimentswithclassicaltexts-2024","role":"author","urls":{"Paper":"https://journal.dhbenelux.org/wp-content/uploads/2024/11/9_Soffiantini_individual.pdf"},"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/zotero-group/schulzkx/5158478","dataSources":["JFDnASMkoQCjjGL8E"],"keywords":[],"search_terms":["cross","linguistic","annotation","transfer","geoparsing","experiments","classical","texts","soffiantini"],"title":"Cross-linguistic annotation transfer in geoparsing experiments with Classical texts","year":2024}