Merge pull request #6725 from JohnSnowLabs/josejuanmartinez-patch-re_drugprot

Update 2022-01-05-re_drugprot_clinical_en.md

Author: vkocaman

docs/_posts/luca-martial/2022-01-05-re_drugprot_clinical_en.md

@@ -17,6 +17,8 @@ use_language_switcher: "Python-Scala-Java"
 
 ## Description
 
+NOTE: This model has been improved by a new SOTA, Bert-based, Relation Extraction model, you can find [here](https://nlp.johnsnowlabs.com/2022/01/05/redl_drugprot_biobert_en.html)
+
 Detect interactions between chemical compounds/drugs and genes/proteins using Spark NLP's `RelationExtractionModel()` by classifying whether a specified semantic relation holds between a chemical and gene entities within a sentence or document. The entity labels used during training were derived from the [custom NER model](https://nlp.johnsnowlabs.com/2021/12/20/ner_drugprot_clinical_en.html) created by our team for the [DrugProt corpus](https://zenodo.org/record/5119892). These include `CHEMICAL` for chemical compounds/drugs, `GENE` for genes/proteins and `GENE_AND_CHEMICAL` for entity mentions of type `GENE` and of type `CHEMICAL` that overlap (such as enzymes and small peptides). The relation categories from the [DrugProt corpus](https://zenodo.org/record/5119892) were condensed from 13 categories to 10 categories due to low numbers of examples for certain categories. This merging process involved grouping the `SUBSTRATE_PRODUCT-OF` and `SUBSTRATE` relation categories together and grouping the `AGONIST-ACTIVATOR`, `AGONIST-INHIBITOR` and `AGONIST` relation categories together.
 
 ## Predicted Entities