Patrick P. K. Chan
South China University of Technology
121 Papers
676 Citations
Patrick P. K. Chan is an academic researcher from South China University of Technology. The author has contributed to research in topics: Computer science & Artificial neural network. The author has an hindex of 20, co-authored 114 publications. Previous affiliations of Patrick P. K. Chan include Hong Kong Polytechnic University.
Chat about Author
Papers
Adversarial Feature Selection Against Evasion Attacks
TL;DR: This paper proposes a novel adversary-aware feature selection model that can improve classifier security against evasion attacks, by incorporating specific assumptions on the adversary's data manipulation strategy.
284
Content-based image retrieval using color moment and Gabor texture feature
Zhi-Chun Huang,Patrick P. K. Chan,Wing W. Y. Ng,Daniel S. Yeung +3 more
- 11 Jul 2010
TL;DR: The proposed content-based image retrieval method has higher retrieval accuracy than conventional methods using color and texture features even though its feature vector dimension results in a lower rate than the conventional method.
244
Static detection of Android malware by using permissions and API calls
Patrick P. K. Chan,Wen-Kai Song +1 more
- 13 Jul 2014
TL;DR: A feature set containing the permissions and API calls for Android malware static detection was proposed and it showed that the information of API calls is helpful in recognizing Android malware.
109
MLPNN Training via a Multiobjective Optimization of Training Error and Stochastic Sensitivity
TL;DR: A stochastic sensitivity measure (ST-SM) is proposed to realize a new penalty term for MLPNN training and is adopted as a two-phase Pareto-based multiobjective training algorithm for minimizing both the training error and the ST-SM as biobjective functions.
70
Item Relationship Graph Neural Networks for E-Commerce.
Weiwen Liu,Yin Zhang,Jianling Wang,Yun He,James Caverlee,Patrick P. K. Chan,Daniel S. Yeung,Pheng-Ann Heng +7 more
TL;DR: Wang et al. as discussed by the authors formulated the problem as a multilabel link prediction task and proposed a novel graph neural network-based framework for discovering multiple complex relationships simultaneously, incorporating multihop relationships of products by recursively updating node embeddings using the messages from their neighbors.
61