A Prior of Saliency Based Pruning Algorithms
نویسنده
چکیده
In saliency based pruning algorithms a pruning decision is made based on an ordering of the parameters. We will in this article focus on the fact that this ordering is invariant under certain transformations, and with that knowledge an equivalence class of algorithms is developed all yielding identical prunings. A sub-class is demonstrated to have a simple but sensible interpretation in terms of Bayesian decision theory.
منابع مشابه
Space-Variant Descriptor Sampling for Action Recognition Based on Saliency and Eye Movements
Algorithms using “bag of features”-style video representations currently achieve state-of-the-art performance on action recognition tasks, such as the challenging Hollywood2 benchmark [1,2,3]. These algorithms are based on local spatiotemporal descriptors that can be extracted either sparsely (at interest points) or densely (on regular grids), with dense sampling typically leading to the best p...
متن کاملCompressed-Sampling-Based Image Saliency Detection in the Wavelet Domain
When watching natural scenes, an overwhelming amount of information is delivered to the Human Visual System (HVS). The optic nerve is estimated to receive around 108 bits of information a second. This large amount of information can’t be processed right away through our neural system. Visual attention mechanism enables HVS to spend neural resources efficiently, only on the selected parts of the...
متن کاملGraph-based Visual Saliency Model using Background Color
Visual saliency is a cognitive psychology concept that makes some stimuli of a scene stand out relative to their neighbors and attract our attention. Computing visual saliency is a topic of recent interest. Here, we propose a graph-based method for saliency detection, which contains three stages: pre-processing, initial saliency detection and final saliency detection. The initial saliency map i...
متن کاملUniversal Distribution of Saliencies for Pruning in Layered Neural Networks
A better understanding of pruning methods based on a ranking of weights according to their saliency in a trained network requires further information on the statistical properties of such saliencies. We focus on two-layer networks with either a linear or nonlinear output unit, and obtain analytic expressions for the distribution of saliencies and their logarithms. Our results reveal unexpected ...
متن کاملComparing Adaptive and Non - AdaptiveConnection Pruning With Pure Early
|Neural network pruning methods on the level of individual network parameters (e.g. connection weights) can improve generalization, as is shown in this empirical study. However, an open problem in the pruning methods known today (OBD, OBS, autoprune, epsiprune) is the selection of the number of parameters to be removed in each pruning step (pruning strength). This work presents a pruning method...
متن کامل