update data

README.md

@@ -144,8 +144,8 @@ For the ***label skew*** scenario, we introduce **16** famous datasets:
 - **GTSRB**
 - **Shakespeare**
 - **Stanford Cars**
-- **COVIDx** (chest X-ray images for covid-19)
-- **kvasir** (endoscopic images for gastrointestinal disease detection)
+- **COVIDx**
+- **kvasir**
 
 The datasets can be easily split into **IID** and **non-IID** versions. In the **non-IID** scenario, we distinguish between two types of distribution:
 
@@ -165,7 +165,7 @@ For the ***feature shift*** scenario, we utilize **3** widely used datasets in D
 ### ***real-world*** scenario
 
 For the ***real-world*** scenario, we introduce **5** naturally separated datasets:  
-- **Camelyon17** (tumor tissue patches extracted from breast cancer metastases in lymph node sections, 5 hospitals, 2 labels)  
+- **Camelyon17** (5 hospitals, 2 labels)  
 - **iWildCam** (194 camera traps, 158 labels)  
 - **Omniglot** (20 clients, 50 labels)  
 - **HAR (Human Activity Recognition)** (30 clients, 6 labels)  

docs/data.html

@@ -199,8 +199,8 @@ <h3><em><strong>Label Skew</strong></em> Scenario</h3>
                 <li><strong>GTSRB</strong></li>
                 <li><strong>Shakespeare</strong></li>
                 <li><strong>Stanford Cars</strong></li>
-                <li><strong>COVIDx</strong> (chest X-ray images for covid-19)</li>
-                <li><strong>kvasir</strong> (endoscopic images for gastrointestinal disease detection)</li>
+                <li><strong>COVIDx</strong></li>
+                <li><strong>kvasir</strong></li>
                 </ul>
 
                 <p>The datasets can be easily split into <strong>IID</strong> and <strong>non-IID</strong> versions. In the <strong>non-IID</strong> scenario, we distinguish between two types of distribution:</p>
@@ -225,7 +225,7 @@ <h3><em><strong>Real-World</strong></em> Scenario</h3>
 
                 <p>For the <strong>real-world</strong> scenario, we introduce <strong>5</strong> naturally separated datasets:</p>
                 <ul>
-                <li><strong>Camelyon17</strong> (tumor tissue patches extracted from breast cancer metastases in lymph node sections, 5 hospitals, 2 labels)</li>
+                <li><strong>Camelyon17</strong> (5 hospitals, 2 labels)</li>
                 <li><strong>iWildCam</strong> (194 camera traps, 158 labels)</li>
                 <li><strong>Omniglot</strong> (20 clients, 50 labels)</li>
                 <li><strong>HAR (Human Activity Recognition)</strong> (30 clients, 6 labels, see examples <a href="#exp-har">here</a>)</li>