Parameter-Free Convex Learning through Coin Betting.
AutoML@ICML(2016)
摘要
We present a new parameter-free algorithm for online linear optimization over any Hilbert space. It is theoretically optimal, with regret guarantees as good as with the best possible learning rate. The algorithm is simple and easy to implement. The analysis is given via the adversarial coin-betting game, Kelly betting and the Krichevsky-Trofimov estimator. Applications to obtain parameter-free convex optimization and machine learning algorithms are shown.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要