Some minor changes to the Model Card published this morning

Added institution + developing projects to quick summary; corrected table of contents; added datasets and publications to the "More information" section

README.md

@@ -19,20 +19,20 @@ license: apache-2.0
 
 <!-- Provide a quick summary of what the model is/does. [Optional] -->
 A BERT model trained on three German corpora containing contemporary and historical texts for named entity recognition tasks. It predicts the classes PER, LOC and ORG. 
-Questions and comments about the model can be directed to Clemens Neudecker at clemens.neudecker@sbb.spk-berlin.de.
-
+The model was developed by the Berlin State Library (SBB) in the [QURATOR](https://staatsbibliothek-berlin.de/die-staatsbibliothek/projekte/project-id-1060-2018) 
+and [Mensch.Maschine.Kultur]( https://mmk.sbb.berlin/?lang=en) projects. 
 
 
 
 #  Table of Contents
 
-- [Model Card for sbb_ner](#model-card-for--model_id-)
+- [Model Card for sbb_ner](#model-card-for-sbb_ner)
 - [Table of Contents](#table-of-contents)
 - [Model Details](#model-details)
   - [Model Description](#model-description)
 - [Uses](#uses)
   - [Direct Use](#direct-use)
-  - [Downstream Use [Optional]](#downstream-use-optional)
+  - [Downstream Use [Optional]](#downstream-use)
   - [Out-of-Scope Use](#out-of-scope-use)
 - [Bias, Risks, and Limitations](#bias-risks-and-limitations)
   - [Recommendations](#recommendations)
@@ -70,13 +70,13 @@ Questions and comments about the model can be directed to Clemens Neudecker at c
 A BERT model trained on three German corpora containing contemporary and historical texts for named entity recognition tasks. 
 It predicts the classes PER, LOC and ORG. 
 
-- **Developed by:** [Kai Labusch](https://huggingface.co/labusch), [Clemens Neudecker](https://huggingface.co/cneud), David Zellhöfer
+- **Developed by:** [Kai Labusch](kai.labusch@sbb.spk-berlin.de), [Clemens Neudecker](clemens.neudecker@sbb.spk-berlin.de), David Zellhöfer
 - **Shared by [Optional]:** [Staatsbibliothek zu Berlin / Berlin State Library](https://huggingface.co/SBB)
 - **Model type:** Language model
 - **Language(s) (NLP):** de
 - **License:** apache-2.0
-- **Parent Model:** The BERT base multilingual cased model as provided by [Google] (https://huggingface.co/bert-base-multilingual-cased)
-- **Resources for more information:** More information needed
+- **Parent Model:** The BERT base multilingual cased model as provided by [Google](https://huggingface.co/bert-base-multilingual-cased)
+- **Resources for more information:** 
     - [GitHub Repo](https://github.com/qurator-spk/sbb_ner)
     - [Associated Paper](https://konvens.org/proceedings/2019/papers/KONVENS2019_paper_4.pdf)
 
@@ -87,7 +87,7 @@ It predicts the classes PER, LOC and ORG.
 
 ## Direct Use
 
-The model can directly be used to perform NER on historical german texts obtained by OCR from digitized documents.
+The model can directly be used to perform NER on historical German texts obtained by OCR from digitized documents.
 Supported entity types are PER, LOC and ORG. 
 
 <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
@@ -95,17 +95,22 @@ Supported entity types are PER, LOC and ORG.
 
 ## Downstream Use
 
-The model has been pre-trained on 2.300.000 pages of OCR-text of the digitized collections of Berlin State Library.
-Therefore it is adapted to OCR-error prone historical german texts and might be used for particular applications that involve such text material.
-
 <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
 <!-- If the user enters content, print that. If not, but they enter a task in the list, use that. If neither, say "more info needed." -->
+ 
+The model has been pre-trained on 2.300.000 pages of OCR-text of the digitized collections of Berlin State Library.
+Therefore it is adapted to OCR-error prone historical German texts and might be used for particular applications that involve such text material.
+
+
+
 
 ## Out-of-Scope Use
 
 <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
 <!-- If the user enters content, print that. If not, but they enter a task in the list, use that. If neither, say "more info needed." -->
 
+More info needed.
+
 
 # Bias, Risks, and Limitations
 
@@ -113,6 +118,7 @@ Therefore it is adapted to OCR-error prone historical german texts and might be
 
 The identification of named entities in historical and contemporary texts is a task contributing to knowledge creation aiming at enhancing scientific research and better discoverability of information in digitized historical texts. The aim of the development of this model was to improve this knowledge creation process, an endeavour that is not for profit. The results of the applied model are freely accessible for the users of the digital collections of the Berlin State Library. Against this backdrop, ethical challenges cannot be identified. As a limitation, it has to be noted that there is a lot of performance to gain for historical text by adding more historical ground-truth data. 
 
+
 ## Recommendations
 
 <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
@@ -131,15 +137,16 @@ The general observation that historical texts often remain silent and avoid nami
 3) DC-SBB Digital Collections of the Berlin State Library (Labusch and Zellhöfer, 2019)
 4) Europeana Newspapers Historic German Datasets (Neudecker, 2016)
 
+
 ## Training Procedure
 
 <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
 
-The BERT model is trained directly with respect to the NER by implementation of the same method that has been proposed by the BERT authors (Devlin et al., 2018). We applied unsupervised pre-training on 2,333,647 pages of unlabeled historical German text from the Berlin State Library digital collections, and supervised pre-training on two datasets with contemporary German text, conll2003 and germeval_14. Unsupervised pre-training on the DC-SBB data as well as supervised pre-training on contemporary NER ground truth were applied. Unsupervised and supervised pretraining are combined where unsupervised is done first and supervised second. Performance on different combinations of training and test sets was explored, and a 5-fold cross validation and comparison with state of the art approaches was conducted.
+The BERT model is trained directly with respect to the NER by implementation of the same method that has been proposed by the BERT authors (Devlin et al., 2018). We applied unsupervised pre-training on 2,333,647 pages of unlabeled historical German text from the Berlin State Library digital collections, and supervised pre-training on two datasets with contemporary German text, conll2003 and germeval_14. Unsupervised pre-training on the DC-SBB data as well as supervised pre-training on contemporary NER ground truth were applied. Unsupervised and supervised pre-training are combined where unsupervised is done first and supervised second. Performance on different combinations of training and test sets was explored, and a 5-fold cross validation and comparison with state of the art approaches was conducted.
 
 ### Preprocessing
 
-The model was pretrained on 2.300.000 pages of german texts from the digitized collections of the Berlin State Library.
+The model was pre-trained on 2.300.000 pages of German texts from the digitized collections of the Berlin State Library.
 The texts have been obtained by OCR from the page scans of the documents.
 
 ### Speeds, Sizes, Times
@@ -147,12 +154,12 @@ The texts have been obtained by OCR from the page scans of the documents.
 <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
 
 Since it is an incarnation of the original BERT-model published by Google, all the speed, size and time considerations of that original model hold.
-
  
 # Evaluation
 
 <!-- This section describes the evaluation protocols and provides the results. -->
-The model has been evaluated by 5-fold cross-validation on several german historical OCR ground truth datasets. 
+
+The model has been evaluated by 5-fold cross-validation on several German historical OCR ground truth datasets. 
 See publication for detail.
 
 ## Testing Data, Factors & Metrics
@@ -161,15 +168,15 @@ See publication for detail.
 
 <!-- This should link to a Data Card if possible. -->
 
-Two different test sets contained in the CoNLL 2003 German Named Entity Recognition Ground Truth, 
-i.e. TEST-A and TEST-B, have been used for testing (DE-CoNLL-TEST).
-Additionaly historical OCR-based ground truth datasets have been used for testing - see publication for details.
+Two different test sets contained in the CoNLL 2003 German Named Entity Recognition Ground Truth, i.e. TEST-A and TEST-B, have been used for testing (DE-CoNLL-TEST).
+Additionally, historical OCR-based ground truth datasets have been used for testing - see publication for details and below.
+
 
 ### Factors
 
 <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
 
-The evaluation focuses on NER in historical germans documents, see publication for details.
+The evaluation focuses on NER in historical German documents, see publication for details.
 
 ### Metrics
 
@@ -182,12 +189,10 @@ See paper for actual results in terms of these metrics.
 
 See publication.
 
-
 # Model Examination
 
 See publication.
 
-
 # Environmental Impact
 
 <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
@@ -195,12 +200,11 @@ See publication.
 Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
 
 - **Hardware Type:** V100
-- **Hours used:** Roughly 1-2 week(s) for pretraining. Roughly 1 hour for final NER-training.
+- **Hours used:** Roughly 1-2 week(s) for pre-training. Roughly 1 hour for final NER-training.
 - **Cloud Provider:** No cloud.
 - **Compute Region:** Germany.
 - **Carbon Emitted:** More information needed
 
-
 # Technical Specifications [optional]
 
 ## Model Architecture and Objective
@@ -217,8 +221,7 @@ See above.
 
 ### Software
 
-See published code on github.
-
+See published code on [GithHub]( https://github.com/qurator-spk/sbb_ner).
 
 # Citation
 
@@ -241,36 +244,37 @@ See published code on github.
 
 (Labusch et al., 2019)
 
-
 # Glossary [optional]
 
 <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
 
-[More information needed]
-
+More information needed.
 
 # More Information [optional]
 
-[More information needed]
+In addition to what has been documented above, it should be noted that there are two NER Ground Truth datasets available:
+
+1) [Data provided for the 2020 HIPE campaign on named entity processing]( https://impresso.github.io/CLEF-HIPE-2020/)
+2) [Data providided for the 2022 HIPE shared task on named entity processing]( https://hipe-eval.github.io/HIPE-2022/) 
+ 
+Furthermore, two papers have been published on NER/NED, using BERT:
+
+1) [Entity Linking in Multilingual Newspapers and Classical Commentaries with BERT]( http://ceur-ws.org/Vol-3180/paper-85.pdf)
+2) [Named Entity Disambiguation and Linking Historic Newspaper OCR with BERT]( http://ceur-ws.org/Vol-2696/paper_163.pdf)
 
 
 # Model Card Authors [optional]
 
 <!-- This section provides another layer of transparency and accountability. Whose views is this model card representing? How many voices were included in its construction? Etc. -->
 
-[Kai Labusch](kai.labusch@sbb.spk-berlin.de) and [Jörg Lehmann](https://huggingface.co/Jrglmn)
+[Kai Labusch](kai.labusch@sbb.spk-berlin.de) and [Jörg Lehmann](joerg.lehmann@sbb.spk-berlin.de)
 
 
 # Model Card Contact
 
-Questions and comments about the model can be directed to Clemens Neudecker at clemens.neudecker@sbb.spk-berlin.de, 
-questions and comments about the model card can be directed to Jörg Lehmann at joerg.lehmann@sbb.spk-berlin.de
-
+Questions and comments about the model can be directed to Clemens Neudecker at clemens.neudecker@sbb.spk-berlin.de, questions and comments about the model card can be directed to Jörg Lehmann at joerg.lehmann@sbb.spk-berlin.de
 
 # How to Get Started with the Model
 
-Use the code below to get started with the model.
 
-<details>
 How to get started with this model is explained in the ReadMe file of the GitHub repository [over here](https://github.com/qurator-spk/sbb_ner).
-</details>