Lei Ren

VLT: A Vision-Language-Time Series Multimodal Foundation Model for Industrial Intelligence

This work proposes VLT, a multimodal foundation model that jointly models time-series, frequency-spectrum visual representations, and textual knowledge, and demonstrates that VLT outperforms state-of-the-art methods, achieving superior robustness and generalization under few-shot, noisy, and incomplete-modality setting...