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

The success of the application of machine-learning techniques to compilationtasks can be largely attributed to the recent development and advancement ofprogram characterization, a process that numerically or structurally quantifiesa target program. While great achievements have been made in identifying keyfeatures to characterize programs, choosing a correct set of features for aspecific compiler task remains an ad hoc procedure. In order to guarantee acomprehensive coverage of features, compiler engineers usually need to selectexcessive number of features. This, unfortunately, would potentially lead to aselection of multiple similar features, which in turn could create a newproblem of bias that emphasizes certain aspects of a program's characteristics,hence reducing the accuracy and performance of the target compiler task. Inthis paper, we propose FEAture Selection for compilation Tasks (FEAST), anefficient and automated framework for determining the most relevant andrepresentative features from a feature pool. Specifically, FEAST utilizeswidely used statistics and machine-learning tools, including LASSO, sequentialforward and backward selection, for automatic feature selection, and can ingeneral be applied to any numerical feature set. This paper further proposes anautomated approach to compiler parameter assignment for assessing theperformance of FEAST. Intensive experimental results demonstrate that, underthe compiler parameter assignment task, FEAST can achieve comparable resultswith about 18% of features that are automatically selected from the entirefeature pool. We also inspect these selected features and discuss their rolesin program execution.

FEAST: An Automated Feature Selection Framework for Compilation Tasks