Update announcement_aaai.md

_news/announcement_aaai.md

@@ -15,3 +15,9 @@ Hung Ba Ngo*, Doanh C. Bui*, Nhat-Tuong Do-Tran, Choi Tae Jong
 In this study, we formulate the domain adaptation problem as a hierarchical graph of nodes. From this perspective, an individual image is considered a local graph of nodes, and a set of image embeddings is considered a global graph, where edges exist if two nodes belong to the same category.
 
 <img src="https://github.com/user-attachments/assets/7b1c3993-0ffd-4fd0-bb4e-da75692b947c" data-canonical-src="https://caodoanh2001.github.io/assets/img/tcsvt-transformer-based.jpg" width="750" height="300"/>
+
+Some visualizations show that our local graph successfully aggregates nodes of interest, which are discriminative for prediction:
+
+<img src="https://github.com/user-attachments/assets/7f0b0688-6b1a-48cb-9984-dffd75327655" data-canonical-src="https://github.com/user-attachments/assets/7f0b0688-6b1a-48cb-9984-dffd75327655" width="750" height="500"/>
+
+We will provide the source code of HiGDA very soon.