YuCCA Lab

Welcome to the Yu’s Cognitive Computing and Automation Lab (YuCCA Lab) at the University of California, Merced!

Research Focus

The YuCCA Lab, led by Dr. Xiaofan Yu, conducts research on efficient and autonomous AI systems that bring intelligent computing to real-world applications. Our research spans the full stack: from designing efficient AI algorithms and embedded computing platforms to deploying intelligent systems in real-world applications. By co-designing hardware, systems, and AI, we build edge platforms that can perceive, reason, and act under stringent computing, memory, and energy constraints.

Active Projects

Multimodal Large Language Models and Agents for Physical AI

Developing efficient and reliable multimodal LLMs and AI agents that integrate diverse sensor inputs to enable autonomous perception, reasoning, and decision-making in real-world environments. Our research focuses on reducing hallucinations, improving reliability, and enabling efficient on-device intelligence for Physical AI.

Publications: NeurIPSW'26, IROS'26, WACV'26, SenSys'25, MASS'25

Intelligent Embedded Systems for Real-World Applications

Designing intelligent and resource-efficient embedded systems that bring advanced AI capabilities to real-world applications. Our research integrates on-device learning, multimodal sensing, and collaborative intelligence to enable autonomous robotic systems, intelligent wearables, and real-time monitoring in resource-constrained environments.

Publications: HumanSys'26, IMWUT'25, IoTDI'23

Hyperdimensional Computing: Algorithms, Systems and Hardware

Advancing neuromorphic and brain-inspired computing paradigms, particularly Hyperdimensional Computing (HDC), for next-generation AI systems. Our research bridges HDC and real-world cyber-physical applications through full-stack solutions spanning algorithms, systems, and emerging hardware, including Processing-in-Memory (PIM).

Publications: DAC'26, AAAI'25, DATE'25, TIOT'25, IPSN'24, DATE'24, ASP-DAC'24

Recent News