Dynamically Scaled Temperature in Self-Supervised Contrastive Learning
This work addresses a key hyperparameter tuning issue in self-supervised learning for computer vision, but it is incremental as it builds on established contrastive methods like SimCLR and MoCo.
The paper tackles the problem of balancing attraction and repulsion in self-supervised contrastive learning by proposing a dynamically scaled temperature function for the InfoNCE loss, resulting in improved performance over existing algorithms.
In contemporary self-supervised contrastive algorithms like SimCLR, MoCo, etc., the task of balancing attraction between two semantically similar samples and repulsion between two samples of different classes is primarily affected by the presence of hard negative samples. While the InfoNCE loss has been shown to impose penalties based on hardness, the temperature hyper-parameter is the key to regulating the penalties and the trade-off between uniformity and tolerance. In this work, we focus our attention on improving the performance of InfoNCE loss in self-supervised learning by proposing a novel cosine similarity dependent temperature scaling function to effectively optimize the distribution of the samples in the feature space. We also provide mathematical analyses to support the construction of such a dynamically scaled temperature function. Experimental evidence shows that the proposed framework outperforms the contrastive loss-based SSL algorithms.