[Mt-list] CfP: Workshop on Multilingual de-identification of (sensitive) LRs at LREC 2022 - Marseille, June 20, 2022
[Apologies for multiple postings] CALL FOR PAPERS Workshop on Multilingual de-identification of (sensitive) language resources To be held in conjunction with the 13th International Language Resources and Evaluation Conference (LREC 2022) 20 June 2022, Le Palais du Pharo, Marseille, France *Deadline for submission: 10 April 2022* *Description* The General Data Protection Regulation (GDPR - Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016) ensures the protection of natural persons with regard to the processing of personal data and on the free movement of such data. The GDPR outlines a specific set of rules that protect citizens and user data and create transparency in information sharing. GDPR is the strictest data privacy regulation in the world, and considerable work is taking place to develop techniques and deploy systems that help comply with this regulation while rendering data accessible and, thus, usable for further processing. Different techniques are studied to guarantee such compliance, implying different levels of sensitive content protection and with a short- or long-term guarantee depending on whether we may have access to additional related information. In this regard, we can read about work on anonymization, de-identification and pseudonymization. While anonymization implies a zero re-identification risk, which is extremely difficult to secure, de-identification and pseudonymization represent an attainable target under the GDPR, given that this regulation defines pseudonymization as “the processing of personal data in such a manner that the personal data can no longer be attributed to a specific data subject without the use of additional information, provided that such additional information is kept separately and is subject to technical and organisational measures to ensure that the personal data are not attributed to an identified or identifiable natural person.” Bearing this context in mind, multilingual approaches and kits for (sensitive) language resources de-identification may provide the means to share language data while also protecting private or sensitive data by spotting then deleting, obfuscating, pseudonymizing or encrypting person identifying information. De-identification is typically performed for the purpose of protecting an individual’s private ctivities while maintaining the usefulness of the gathered data for research and development purposes. This workshop aims at discussing the various approaches to effective and reliable text de-identification, focusing on some sensitive domains such as the medical and legal domains, but not only. Based on these premises a consensus emerges that shows a clear situation and needs: 1. Tools for the multilingual de-identification of (sensitive) language resources are becoming essential to ensure that such resources can be shared. 2. De-identification is crucial to ensure that all legal & ethical considerations are taken into account during the production/repurposing phases but also that the quality/nature of the de-identified data sets remains appropriate to conduct research activities. 3. European Public Administrations need personal data processing tools to handle the extremely large amounts of data they manage. 4. Europe’s multilingual context will benefit from approaches and tools that can support the European Digital Market in their multilingual data exchanges. *Workshop Objectives and Topics of Interest** * This workshop is organised by members of the MAPA project, funded by the EU Connecting Europe Facility (CEF) program (https://mapa-project.eu/). This project has developed a toolkit for the de-identification of texts in the medical and legal fields which addresses all EU official languages. It has followed a BERT-based Named Entity Recognition approach for personal information identification. A wide range of topics have been considered and are hot topics open for discussion to all participants of this workshop. Among them, we have the following: 1. Sensitive personal information, domains and services that require de-identification 2. Corpora annotation and/or creation 3. Annotation guidelines and platforms 4. De-identification tools, data and/or applications 5. De-identification and minority languages 6. Multi-domain and/or multilingual processing 7. NLP techniques and tools used for de-identification 8. Multimodal de-identification 9. Validation and benchmarking of de-identified resources 10. Evaluation of de-identification tools and applications 11. Evaluation protocols: how to evaluate, metrics, approaches, data, experiences 12. Best practices 13. Approaches, activities and systems addressing “anonymization” are also welcome to share their experience. 14. Any other topic related to de-identification This workshop will also be a good forum to discuss the possibility to design and initiate a new
[Mt-list] CfP: Workshop on Multilingual de-identification of (sensitive) LRs at LREC 2022 - Marseille, June 20, 2022
[Apologies for multiple postings] CALL FOR PAPERS Workshop on Multilingual de-identification of (sensitive) language resources To be held in conjunction with the 13th International Language Resources and Evaluation Conference (LREC 2022) 20 June 2022, Le Palais du Pharo, Marseille, France *Deadline for submission: 10 April 2022* *Description* The General Data Protection Regulation (GDPR - Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016) ensures the protection of natural persons with regard to the processing of personal data and on the free movement of such data. The GDPR outlines a specific set of rules that protect citizens and user data and create transparency in information sharing. GDPR is the strictest data privacy regulation in the world, and considerable work is taking place to develop techniques and deploy systems that help comply with this regulation while rendering data accessible and, thus, usable for further processing. Different techniques are studied to guarantee such compliance, implying different levels of sensitive content protection and with a short- or long-term guarantee depending on whether we may have access to additional related information. In this regard, we can read about work on anonymization, de-identification and pseudonymization. While anonymization implies a zero re-identification risk, which is extremely difficult to secure, de-identification and pseudonymization represent an attainable target under the GDPR, given that this regulation defines pseudonymization as “the processing of personal data in such a manner that the personal data can no longer be attributed to a specific data subject without the use of additional information, provided that such additional information is kept separately and is subject to technical and organisational measures to ensure that the personal data are not attributed to an identified or identifiable natural person.” Bearing this context in mind, multilingual approaches and kits for (sensitive) language resources de-identification may provide the means to share language data while also protecting private or sensitive data by spotting then deleting, obfuscating, pseudonymizing or encrypting person identifying information. De-identification is typically performed for the purpose of protecting an individual’s private ctivities while maintaining the usefulness of the gathered data for research and development purposes. This workshop aims at discussing the various approaches to effective and reliable text de-identification, focusing on some sensitive domains such as the medical and legal domains, but not only. Based on these premises a consensus emerges that shows a clear situation and needs: 1. Tools for the multilingual de-identification of (sensitive) language resources are becoming essential to ensure that such resources can be shared. 2. De-identification is crucial to ensure that all legal & ethical considerations are taken into account during the production/repurposing phases but also that the quality/nature of the de-identified data sets remains appropriate to conduct research activities. 3. European Public Administrations need personal data processing tools to handle the extremely large amounts of data they manage. 4. Europe’s multilingual context will benefit from approaches and tools that can support the European Digital Market in their multilingual data exchanges. *Workshop Objectives and Topics of Interest** * This workshop is organised by members of the MAPA project, funded by the EU Connecting Europe Facility (CEF) program (https://mapa-project.eu/). This project has developed a toolkit for the de-identification of texts in the medical and legal fields which addresses all EU official languages. It has followed a BERT-based Named Entity Recognition approach for personal information identification. A wide range of topics have been considered and are hot topics open for discussion to all participants of this workshop. Among them, we have the following: 1. Sensitive personal information, domains and services that require de-identification 2. Corpora annotation and/or creation 3. Annotation guidelines and platforms 4. De-identification tools, data and/or applications 5. De-identification and minority languages 6. Multi-domain and/or multilingual processing 7. NLP techniques and tools used for de-identification 8. Multimodal de-identification 9. Validation and benchmarking of de-identified resources 10. Evaluation of de-identification tools and applications 11. Evaluation protocols: how to evaluate, metrics, approaches, data, experiences 12. Best practices 13. Approaches, activities and systems addressing “anonymization” are also welcome to share their experience. 14. Any other topic related to de-identification This workshop will also be a good forum to discuss the possibility to design and initiate a new