arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2211.09875stat.CO
收藏

专家混合分布回归:利用鲁棒估计与自适应一阶方法实现

Mixture of Experts Distributional Regression: Implementation Using Robust Estimation with Adaptive First-order Methods

LMU Munich(慕尼黑大学);TU Dortmund(多特蒙德工业大学);Munich Center for Machine Learning(慕尼黑机器学习中心);WU Vienna(维也纳经济大学)

David Rügamer, Florian Pfisterer, Bernd Bischl, Bettina Grün

AI总结:

本文提出了一种高效的专家混合分布回归模型实现方法,利用随机一阶优化技术与自适应学习率调度器进行鲁棒估计,通过混合distreg R包在高维和大样本设置下实现多种分布族的混合估计。

摘要:

本文提出了一种高效的专家混合分布回归模型实现方法,该方法利用随机一阶优化技术与自适应学习率调度器进行鲁棒估计。我们利用神经网络软件的灵活性和可扩展性,在mixdistreg R软件包中实现了该框架,允许定义多种分布族的混合,支持高维和大样本设置下的估计,并基于TensorFlow实现鲁棒优化。数值实验显示,优化的可靠性与经典方法在多种设置下相当,并能在经典方法持续失败的复杂场景中获得结果。

英文摘要:

In this work, we propose an efficient implementation of mixtures of experts distributional regression models which exploits robust estimation by using stochastic first-order optimization techniques with adaptive learning rate schedulers. We take advantage of the flexibility and scalability of neural network software and implement the proposed framework in mixdistreg, an R software package that allows for the definition of mixtures of many different families, estimation in high-dimensional and large sample size settings and robust optimization based on TensorFlow. Numerical experiments with simulated and real-world data applications show that optimization is as reliable as estimation via classical approaches in many different settings and that results may be obtained for complicated scenarios where classical approaches consistently fail.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑