Chronicle: A Multimodal Foundation Model for Joint Language and Time Series Understanding
Researchers have introduced Chronicle, a novel 324 million parameter transformer model designed for joint understanding of natural language and time series data. Unlike previous approaches that adapt existing language models, Chronicle was trained from scratch with a unified architecture, allowing both modalities to share transformer blocks and attention mechanisms. This approach enables cross-modal capabilities to emerge organically from shared parameters. Chronicle demonstrates competitive performance, matching Gemma-3-270M on natural language understanding tasks and setting new benchmarks for time series classification and multimodal forecasting. AI
IMPACT Introduces a novel architecture for processing combined text and time series data, potentially improving forecasting and analytical capabilities in domains with rich textual context.