Multilingual NER Toolkit: Difference between revisions

Content deleted Content added

Inline

@@ Line 7: / Line 7: @@
             "statements": {
                 "Description": {
-                    "type": "text",
+                    "propertyType": "text",
                     "value": [
                         "Open-source named-entity recognition tooling tuned for historical European languages."
@@ Line 13: / Line 13: @@
                 },
                 "Started": {
-                    "type": "number",
+                    "propertyType": "number",
                     "value": 2022
                 },
                 "Lead institution": {
-                    "type": "relation",
+                    "propertyType": "relation",
                     "value": [
                         {

Latest revision as of 22:06, 27 July 2026

The Multilingual NER Toolkit is an open-source named-entity recognition project tuned for historical European languages. Started in 2022, it ships models and tooling that handle the spelling variation, archaic vocabulary, and code-switching common in pre-modern texts.

Led by Sorbonne University, it serves as a foundation for downstream work on entity disambiguation and linking across cultural heritage corpora.