Problem
Research questions and friction points this paper is trying to address.
energy efficiency
artificial intelligence
neuromorphic computing
CMOS limitations
compute-in-memory
Neuromorphic Computing for Low-Power Artificial Intelligence
🤖 AI Summary
This work addresses the fundamental energy-efficiency limitations of conventional CMOS computing, which struggles to meet the ever-increasing computational and memory demands of artificial intelligence. To overcome this challenge, the study proposes a holistic neuromorphic computing paradigm that integrates novel materials, non-volatile devices, mixed-signal circuits, and physics-aware learning algorithms. By co-designing across multiple abstraction layers—leveraging in-memory computing, analog computation, sparse event-driven communication, and hardware-aware training—the proposed architecture achieves significant improvements in both energy efficiency and scalability. Systematic evaluations demonstrate that this full-stack approach provides a viable theoretical foundation and practical pathway toward next-generation low-power intelligent systems.
📝 Abstract
Classical computing is beginning to encounter fundamental limits of energy efficiency. This presents a challenge that can no longer be solved by strategies such as increasing circuit density or refining standard semiconductor processes. The growing computational and memory demands of artificial intelligence (AI) require disruptive innovation in how information is represented, stored, communicated, and processed. By leveraging novel device modalities and compute-in-memory (CIM), in addition to analog dynamics and sparse communication inspired by the brain, neuromorphic computing offers a promising path toward improvements in the energy efficiency and scalability of current AI systems. But realizing this potential is not a matter of replacing one chip with another; rather, it requires a co-design effort, spanning new materials and non-volatile device structures, novel mixed-signal circuits and architectures, and learning algorithms tailored to the physics of these substrates. This article surveys the key limitations of classical complementary metal-oxide-semiconductor (CMOS) technology and outlines how such cross-layer neuromorphic approaches may overcome them.
Innovation
Methods, ideas, or system contributions that make the work stand out.
neuromorphic computing
compute-in-memory
analog dynamics
sparse communication
co-design
🔎 Similar Papers
No similar papers found.
