Stochastic Non-convex Ordinal Embedding with Stabilized Barzilai-Borwein Step Size
This work addresses the problem of scaling ordinal embedding for large datasets, offering a more efficient solution for researchers and practitioners in machine learning dealing with similarity comparisons, though it is incremental as it builds on existing stochastic and optimization methods.
The paper tackles the computational inefficiency of existing batch methods for ordinal embedding, which rely on convex optimization and SVD, by proposing a stochastic algorithm called SVRG-SBB that is SVD-free and uses an adaptive step size, achieving convergence to a stationary point at a rate of O(1/T) and demonstrating much lower computational cost with good prediction performance in experiments.
Learning representation from relative similarity comparisons, often called ordinal embedding, gains rising attention in recent years. Most of the existing methods are batch methods designed mainly based on the convex optimization, say, the projected gradient descent method. However, they are generally time-consuming due to that the singular value decomposition (SVD) is commonly adopted during the update, especially when the data size is very large. To overcome this challenge, we propose a stochastic algorithm called SVRG-SBB, which has the following features: (a) SVD-free via dropping convexity, with good scalability by the use of stochastic algorithm, i.e., stochastic variance reduced gradient (SVRG), and (b) adaptive step size choice via introducing a new stabilized Barzilai-Borwein (SBB) method as the original version for convex problems might fail for the considered stochastic \textit{non-convex} optimization problem. Moreover, we show that the proposed algorithm converges to a stationary point at a rate $\mathcal{O}(\frac{1}{T})$ in our setting, where $T$ is the number of total iterations. Numerous simulations and real-world data experiments are conducted to show the effectiveness of the proposed algorithm via comparing with the state-of-the-art methods, particularly, much lower computational cost with good prediction performance.