Update README.md

README.md

@@ -10,9 +10,10 @@ size_categories:
 - 10K<n<100K
 pretty_name: MedMKG
 ---
-# MedMKG: Medical Multimodal Knowledge Graph
 
-We introduce **MedMKG**, a **Med**ical **M**ultimodal **K**nowledge **G**raph that seamlessly fuses clinical concepts with medical images.  
+# RADM: Radiological Multimodal Knowledge Graph
+
+We introduce **RADM**, a a **Rad**iological **M**ultimodal Knowledge Graph that seamlessly fuses clinical concepts with medical images.  
 MedMKG is constructed via a multi-stage pipeline that accurately identifies and disambiguates medical concepts while extracting their interrelations.  
 To ensure the conciseness of the resulting graph, we further employ a pruning strategy based on our novel Neighbor-aware Filtering (NaF) algorithm.
 
@@ -21,7 +22,7 @@ To ensure the conciseness of the resulting graph, we further employ a pruning st
 ## 📂 Provided Files
 
 This repository contains:
-- `knowledge_graph.csv` — biomedical triplets: Head, Relation, Tail, Head_Name, Tail_Name
+- `RADM.csv` — biomedical triplets: Head, Relation, Tail, Head_Name, Tail_Name
 - `image_mapping.csv` — image ID to **relative path** mappings
 
 **Note:** The images themselves are **not included**. Users must download MIMIC-CXR-JPG separately and specify their local path.
@@ -57,11 +58,10 @@ To use the image data, you **must** request access and agree to the data use agr
 Below is a demo script to load and link the knowledge graph with your local image data:
 
 ```python
-from huggingface_hub import hf_hub_download
 import pandas as pd
 
-kg_path = hf_hub_download(repo_id="xcwangpsu/MedMKG", filename="MedMKG.csv", repo_type="dataset")
-mapping_path = hf_hub_download(repo_id="xcwangpsu/MedMKG", filename="image_mapping.csv", repo_type="dataset")
+kg_path = "RADM.csv"
+mapping_path = "image_mapping.csv"
 
 # Load CSVs
 kg_df = pd.read_csv(kg_path)
@@ -82,3 +82,15 @@ kg_df["Tail_Path"] = kg_df["Tail"].map(iid_to_path)
 
 print(kg_df.head())
 
+
+```
+
+## Benchmark Instruction
+
+We cover codes for three downstream tasks detailed in our paper, including link prediction, knowledge-augmented visual question answering, and knowledge-augmented text-image retrieval. You may check the three folders and execute the following:
+
+```bash
+python main.py
+```
+
+after specifying your local paths of data files.