Department of Mathematics, School of Transportation Science and Engineering, Harbin Institute of Technology, Harbin 150090, China
Mini Review
Managing Data Bottlenecks: Strategies for Efficient Data Flow across Bandwidth, Storage and Processing
Author(s): Yaping Zhang*
Algorithmically, DMI-ideal arrangements can be inferred by means of the Discriminant Part Investigation (DCA). In addition, DCA has two machine
learning variants that are suitable for supervised learning applications—one in the kernel space and the other in the original space. CP unifies the
conventional Information Bottleneck (IB) and Privacy Funnel (PF) and results in two constrained optimizers known as Generalized Information
Bottleneck (GIB) and Generalized Privacy Funnel (GPF) by extending the concept of DMI to the utility gain and privacy loss. DCA can be further
extended to a DUCA machine learning variant in supervised learning environments to achieve the best possible compromise between utility gain
and privacy loss. Finally, a golden-section iterative method is developed specifically for the two constrained optimization problems in order to speed
.. Read More»
DOI:
10.37421/1736-4337.2024.18.430
Spanish 
Chinese 
Russian 
German 
French 
Japanese 
Portuguese 
Hindi 