English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Kernel approximation is widely used to scale up kernel SVM training andprediction. However, the memory and computation costs of kernel approximationmodels are still too high if we want to deploy them on memory-limited devicessuch as mobile phones, smartwatches, and IoT devices. To address thischallenge, we propose a novel memory and computation-efficient kernel SVM modelby using both binary embedding and binary model coefficients. First, we proposean efficient way to generate compact binary embedding of the data, preservingthe kernel similarity. Second, we propose a simple but effective algorithm tolearn a linear classification model with ternary coefficients that can supportdifferent types of loss function and regularizer. Our algorithm can achievebetter generalization accuracy than existing works on learning binarycoefficients since we allow coefficient to be $-1$, $0$, or $1$ during thetraining stage, and coefficient $0$ can be removed during model inference forbinary classification. Moreover, we provide a detailed analysis of theconvergence of our algorithm and the inference complexity of our model. Theanalysis shows that the convergence to a local optimum is guaranteed, and theinference complexity of our model is much lower than other competing methods.Our experimental results on five large real-world datasets have demonstratedthat our proposed method can build accurate nonlinear SVM models with memorycosts less than 30KB.

Memory and Computation-Efficient Kernel SVM via Binary Embedding and Ternary Model Coefficients