InertialAI

Research at InertialAI

Our research explores how foundation models can understand time-series, sensor data, metadata, and natural language in one system. This page will collect papers, benchmarks, and technical reports as they are released.

Publications

Chronicle is currently available as a preprint. We will replace the local copy with the arXiv version once it is live.

Chronicle paper first page
Preprint

Chronicle: A Multimodal Foundation Model for Joint Language and Time Series Understanding

A compact multimodal transformer trained from scratch on natural language and time-series data with a shared backbone.

Paul Quinlan, Jeremy Levasseur, Qingguo Li, Xiaodan Zhu
Open PDF