Xiangzhong Luo

The paper investigates how LLMs allocate their internal computational depth during multi-turn agentic planning, finding that agents progressively recruit deeper layers and shift toward corrective updates as reasoning complexity increases.

This paper proposes a dynamic image cropping framework and compound shrinking strategy to reduce computational overhead of CNNs on edge AI, achieving higher accuracy and lower computational cost.

This paper addresses the obstacles of using Crossbar-based In-Memory Processing (IMP) accelerators for deep neural networks (DNNs) by reusing bit-shift units for multiplication, applying pruning methods, and adapting to non-ideality.