Session

DescriptionAs AI continues to push toward real-time and resource-efficient processing, optimizing both compute and memory across diverse hardware platforms becomes crucial. This session introduces techniques that aim to improve AI efficiency at the edge, including federated learning frameworks that account for real-world constraints, proactive strategies for mitigating inference cold starts, and novel data streaming techniques that decouple memory access. Additionally, new approaches in cross-layer simulation, task-oriented detection, and low-latency graph processing on FPGAs showcase how hardware-software co-design can unlock smarter, faster, and more efficient AI systems.