BriefGPT.xyz
Feb, 2019
利用平滑凸代理实现结构化预测的一般理论
A General Theory for Structured Prediction with Smooth Convex Surrogates
HTML
PDF
Alex Nowak-Vila, Francis Bach, Alessandro Rudi
TL;DR
本文提出了一个理论框架来进行结构化预测，其将多标签，排序回归和图匹配等任务的损失统一了起来，并可以获得条件随机场和二次代理等现有代理方法上的新结果。
Abstract
In this work we provide a theoretical framework for
structured prediction
that generalizes the existing theory of
surrogate methods
for binary and multiclass classification based on estimating conditional probabi
→