What are the two complementary contrastive modules in TS-TCC and what is the primary purpose of each?

TS-TCC's two complementary contrastive modules are the temporal contrasting module and the contextual contrasting module. The temporal contrasting module captures temporal dependencies by performing a cross-view prediction task, forecasting future features of one augmented view using the historical context of the other. The contextual contrasting module enhances feature discriminability by aligning the contextual embeddings of the two models, maximizing similarity within the same sample and minimizing it across different samples.

The temporal contrasting module is designed to extract temporal dependencies from latent features. It uses autoregressive models to form context vectors from past features, and then performs a cross-view prediction task: the context from the strongly augmented view predicts future features of the weakly augmented view, and vice versa. This challenging task encourages the model to generalize across augmented variants of the same signal and learn robust temporal representations. The contextual contrasting module then operates on the contextual embeddings produced by the two autoregressive models. Its purpose is to enhance feature discriminability by maximizing the similarity between contextual embeddings of the same sample while minimizing similarity across different samples within a batch. Thus, temporal contrasting focuses on capturing temporal structure and robustness, while contextual contrasting focuses on making the learned representations more separable and distinctive.