Online linear optimization with the log-determinant regularizer

TL;DR

This paper introduces a novel online optimization algorithm using log-determinant regularization for positive semi-definite matrices, achieving optimal performance especially with sparse loss matrices, and extends the approach to vector cases.

Contribution

It develops a new performance analysis technique exploiting strong convexity of the log-determinant, improving bounds over previous norm-based methods, and applies to online collaborative filtering.

Findings

Log-determinant regularization outperforms other regularizers with sparse losses.

Achieves optimal guarantees in online collaborative filtering.

Extends the method to vector online optimization with Burg entropy.

Abstract

We consider online linear optimization over symmetric positive semi-definite matrices, which has various applications including the online collaborative filtering. The problem is formulated as a repeated game between the algorithm and the adversary, where in each round t the algorithm and the adversary choose matrices X_t and L_t, respectively, and then the algorithm suffers a loss given by the Frobenius inner product of X_t and L_t. The goal of the algorithm is to minimize the cumulative loss. We can employ a standard framework called Follow the Regularized Leader (FTRL) for designing algorithms, where we need to choose an appropriate regularization function to obtain a good performance guarantee. We show that the log-determinant regularization works better than other popular regularization functions in the case where the loss matrices L_t are all sparse. Using this property, we show that our algorithm achieves an optimal performance guarantee for the online collaborative filtering. The technical contribution of the paper is to develop a new technique of deriving performance bounds by exploiting the property of strong convexity of the log-determinant with respect to the loss matrices, while in the previous analysis the strong convexity is defined with respect to a norm. Intuitively, skipping the norm analysis results in the improved bound. Moreover, we apply our method to online linear optimization over vectors and show that the FTRL with the Burg entropy regularizer, which is the analogue of the log-determinant regularizer in the vector case, works well.