add repo card

README.md

@@ -38,15 +38,24 @@ configs:
     path: data/val-*
 ---
 
----
-{{ dataset_card }}
 ---
 
 # Proposed Active Learning Data split
+## Overview
+This dataset release represents a proposed and experimental data split designed specifically to support and validate planned active learning (AL) cycles for biomedical Named Entity Recognition (NER).
+
+The current version is not a final benchmark split. Instead, it serves as an initial, controlled setup for testing active learning strategies, model uncertainty sampling, 
+and iterative annotation workflows prior to large-scale development.
+
+Both splits (train and validation) have been carefully curated to ensure coverage of all three target entity types:
+
+- **CellLine**
+
+- **CellType**
 
-This version of dataset is just a proposed split to test our planned active learning cycles
-Each of the split(train, validation), have been carefully selected to represent all 3 entities of interest
-**CellLine, CellType and Tissue**.
+- **Tissue**
+
+This balanced representation is critical for meaningful evaluation of active learning behavior across heterogeneous biomedical entities.
 
 
 ##Dataset Description
@@ -56,19 +65,50 @@ Each of the split contain the following features:
 - entities: list of dict of entities found in sentences
 - data_source: the source of the aricle where the sentence orginates
 
-Our dataset have been currated from 3 different domains to represent the 3 entities of interest:
+The dataset has been curated from three complementary biomedical domains, each contributing distinct entity distributions:
 1. Single-Cell transciptomics: rich in CellTypes and Tissues
 2. ChembL assay desriptions: rich in CellLines
 3. Stem-Cell research: contains all 3 entities
 
-The articles from Stem-Cell have been collected from the CellFinder data repository:
-The creation CellFinder dataset is detailed at:
+The stem cell–related articles were collected from the CellFinder data repository.
+The creation and annotation methodology of the original CellFinder dataset are described in the following reference:
 
 >Mariana Neves, Alexander Damaschun, Andreas Kurtz, Ulf Leser (2012)
 >Annotating and evaluating text for stem cell research. 
 >In Proceedings Third Workshop on Building and Evaluation Resources for Biomedical Text Mining (BioTxtM 2012),
 >Language Resources and Evaluation (LREC) 2012.
 
+## Article PMCIDs and source
+### Train Set
+| PMCID | In-Split | Source |
+|-------|---------|--------|
+| PMC12435838 | train | Single-Cell |
+| PMC11578878 | train | Single-Cell |
+| PMC12396968 | train | Single-Cell |
+| PMC11116453 | train | Single-Cell |
+| PMC12408821 | train | Single-Cell |
+| PMC10968586 | train | CheMBL-V1 |
+| PMC10761218 | train | CheMBL-V1 |
+| PMC7642379 | train | CheMBL-V1 |
+| PMC10674574 | train | CheMBL-V1 |
+| PMC1315352 | train | CellFinder |
+| PMC2041973 | train | CellFinder |
+| PMC2238795 | train | CellFinder |
+
+
+### Validation Set
+| PMCID | In-Split | Source |
+|-------|---------|--------|
+| PMC12256823 | val | Single-Cell |
+| PMC12116388 | val | Single-Cell |
+| PMC10287567 | val | Single-Cell |
+| PMC12133578 | val | Single-Cell |
+| PMC8658661 | val | CheMBL-V1 |
+| PMC11350568 | val | CheMBL-V1 |
+| PMC12072392 | val | CheMBL-V2/CeLLaTe-V2 |
+| PMC12115102 | val | CheMBL-V2/CeLLaTe-V2 |
+| PMC2063610 | val | CellFinder |
+| PMC1462997 | val | CellFinder |
 
 ---
 
@@ -82,4 +122,11 @@ The creation CellFinder dataset is detailed at:
 
 ---
 
+## Notes and Limitations
+
+- This is an experimental split, subject to change.
+
+- Entity distributions may not reflect real-world prevalence.
+
+- Annotation density varies across domains by design