Simple and Effective Balance of Contrastive Losses
This work addresses a practical implementation issue in contrastive learning for researchers and practitioners, offering an incremental improvement in optimization efficiency.
The paper tackles the sub-optimal balance of positive and entropy terms in contrastive losses by framing it as a hyper-parameter optimization problem, proposing a coordinate descent-based search method that finds optimal hyper-parameters faster than other methods, as shown in benchmarks from deep metric learning and self-supervised learning.
Contrastive losses have long been a key ingredient of deep metric learning and are now becoming more popular due to the success of self-supervised learning. Recent research has shown the benefit of decomposing such losses into two sub-losses which act in a complementary way when learning the representation network: a positive term and an entropy term. Although the overall loss is thus defined as a combination of two terms, the balance of these two terms is often hidden behind implementation details and is largely ignored and sub-optimal in practice. In this work, we approach the balance of contrastive losses as a hyper-parameter optimization problem, and propose a coordinate descent-based search method that efficiently find the hyper-parameters that optimize evaluation performance. In the process, we extend existing balance analyses to the contrastive margin loss, include batch size in the balance, and explain how to aggregate loss elements from the batch to maintain near-optimal performance over a larger range of batch sizes. Extensive experiments with benchmarks from deep metric learning and self-supervised learning show that optimal hyper-parameters are found faster with our method than with other common search methods.