We are a dynamic research group in the Department of Computer Science and Engineering at UT Arlington (UTA), located at Dallas-Fort Worth Metroplex. Our aim is to build the next-generation efficient computing engine for democratizing artificial intelligence to real-world on-device applications. To reach this goal, we develop the software/hardware co-optimized AI systems based on higher-order tensor decomposition and optimization algorithms.
Specifically, we leverage mathematical tools to solve the efficiency problems fundamentally and improve the task accuracy. We also develop the corresponding hardware/GPU accelerators for our specific software designs.
Our lab is exclusively equipped with multiple high-performance computing servers worth a total of $150K (two 8-Ada6000 servers and two dual-4090 servers). Our department is building a high-performance computing cluster with 6 nodes (each node has eight H100 GPU cards). We are open to collaborate and exchange ideas with external researchers.