AI model learns to predict factory data across different systems without retraining

Researchers have developed an AI system that combines language models with real-time learning to forecast industrial performance across different equipment and factories without being retrained for each new setting. The advance could reduce deployment costs and downtime for manufacturers operating diverse production environments.

Originaltitel: Multimodal knowledge-enhanced language model with online test-time adaptation for cross-domain industrial tabular prediction