Position: Foundation Models Need Digital Twin Representations

TL;DR

This paper advocates for replacing token-based representations in foundation models with digital twin representations to better encode domain knowledge, maintain semantic coherence, and improve reasoning about real-world processes.

Contribution

It introduces digital twin representations as a novel alternative to token-based models for foundation models, emphasizing their advantages in domain knowledge encoding and continuous process modeling.

Findings

Digital twin representations can explicitly encode domain knowledge.

They help preserve the continuous nature of real-world processes.

DTs improve semantic coherence and causal reasoning in FMs.

Abstract

Current foundation models (FMs) rely on token representations that directly fragment continuous real-world multimodal data into discrete tokens. They limit FMs to learning real-world knowledge and relationships purely through statistical correlation rather than leveraging explicit domain knowledge. Consequently, current FMs struggle with maintaining semantic coherence across modalities, capturing fine-grained spatial-temporal dynamics, and performing causal reasoning. These limitations cannot be overcome by simply scaling up model size or expanding datasets. This position paper argues that the machine learning community should consider digital twin (DT) representations, which are outcome-driven digital representations that serve as building blocks for creating virtual replicas of physical processes, as an alternative to the token representation for building FMs. Finally, we discuss how DT representations can address these challenges by providing physically grounded representations that explicitly encode domain knowledge and preserve the continuous nature of real-world processes.