Skip to yearly menu bar Skip to main content


TTOpt: A Maximum Volume Quantized Tensor Train-based Optimization and its Application to Reinforcement Learning

Konstantin Sozykin · Andrei Chertkov · Roman Schutski · Anh-Huy Phan · Andrzej S CICHOCKI · Ivan Oseledets

Keywords: [ Optimization ] [ Reinforcement Learning ] [ Black-box ] [ Quantized networks ] [ Maximum Volume ] [ Tensor Train ] [ Cross approximation ]


We present a novel procedure for optimization based on the combination of efficient quantized tensor train representation and a generalized maximum matrix volume principle.We demonstrate the applicability of the new Tensor Train Optimizer (TTOpt) method for various tasks, ranging from minimization of multidimensional functions to reinforcement learning.Our algorithm compares favorably to popular gradient-free methods and outperforms them by the number of function evaluations or execution time, often by a significant margin.

Chat is not available.