Running large language models on consumer devices such as laptops and desktops is challenging because model weights often exceed GPU memory capacity, making offloading inference necessary to extend effective model capacity with CPU memory. Existing offloading systems, however, ty…
Time-dependent reliability analysis is crucial for ensuring the long-term safety and performance of engineering systems under uncertainties. However, traditional surrogate model methods often struggle to incorporate time-independent random variables and capture their complex inte…
Against the backdrop of global digitalization and green development, digital innovation ecosystems have emerged as key drivers for advancing regional green innovation cooperation and achieving sustainable development goals. This study constructs a theoretical analytical framework…