Time Series, Vision, and Language: Exploring the Limits of Alignment in Contrastive Representation Spaces

TL;DR

This paper investigates the alignment of time series, vision, and language representations in contrastive spaces, revealing asymmetries and the influence of model size and data richness on multimodal alignment.

Contribution

It extends the Platonic Representation Hypothesis to include time series and analyzes the geometric and scaling properties of contrastive alignment across modalities.

Findings

Time series and vision/language encoders are initially orthogonal without coupling.

Alignment improves with model size but is asymmetric, favoring vision over language.

Richer textual descriptions enhance alignment only up to a certain point.

Abstract

The Platonic Representation Hypothesis posits that learned representations from models trained on different modalities converge to a shared latent structure of the world. However, this hypothesis has largely been examined in vision and language, and it remains unclear whether time series participate in such convergence. We first examine this in a trimodal setting and find that independently pretrained time series, vision, and language encoders exhibit near-orthogonal geometry in the absence of explicit coupling. We then apply post-hoc alignment by training projection heads over frozen encoders using contrastive learning, and analyze the resulting representations with respect to geometry, scaling behavior, and dependence on information density and input modality characteristics. Our investigation reveals that overall alignment in contrastive representation spaces improves with model size, but this alignment is asymmetric: time series align more strongly with visual representations than with text, and images can act as effective intermediaries between time series and language. We further see that richer textual descriptions improve alignment only up to a threshold; training on denser captions does not lead to further improvement. Analogous effects are observed for visual representations. Our findings shed light on considerations for building multimodal systems involving non-conventional data modalities beyond vision and language.