GARNET is accepted to LoG'22 as a spotlight paper.

GARNET is accepted to LoG'22 as a spotlight paper.

Chenhui Deng and Prof. Zhiru Zhang, along with their co-authors, Xiuyu Li and Zhuo Feng, will publish their work, GARNET, as a part of this year's proceedings of the Learning on Graphs Conference (LoG).

GARNET is a new method to increase the resilience of graph neural networks against adversarial attacks without impairing the clean graph structures used for training. Their technique shows significant speedups and accuracy improvements over prior work, earning the work a spotlight place in the conference. The recording of Chenhui’s talk can be watched here.

Here is the link to the paper.