Track 1: AI Accelerators & Neuromorphic Computing

With the rapid development of artificial intelligence, particularly large language models, the demand for AI computing power continues to grow, posing severe challenges to the traditional Von Neumann architecture in terms of power consumption, bandwidth, and scalability. Concurrently, emerging computing architectures such as near-memory/processing-in-memory, neuromorphic computing, and heterogeneous multi-chip systems are becoming vital directions to break through AI computing bottlenecks. This forum focuses on the cutting-edge advancements in AI accelerators and neuromorphic computing architectures. It centers around hot topics including AI ASIC accelerators, hardware-software co-design, near-memory and processing-in-memory computing, neuromorphic algorithms and chips, and heterogeneous computing systems, aiming to promote the integrated innovation of novel intelligent computing architectures, chip designs, and system software, as well as to facilitate deep exchanges between academia and industry.

随着人工智能尤其是大模型技术的快速发展，AI算力需求持续增长，传统冯·诺依曼架构在功耗、带宽与扩展性方面面临严峻挑战。与此同时，近存/存内计算、类脑计算、以及异构多芯片系统等新型计算架构，正成为突破AI算力瓶颈的重要方向。本论坛聚焦AI算力芯片与类脑计算架构领域的前沿进展，围绕AI ASIC加速芯片、软硬件协同设计、近存与存内计算、类脑算法与芯片、异构计算系统等热点问题展开交流，旨在推动新型智能计算体系结构、芯片设计与系统软件的融合创新，促进学术界与产业界的深入交流。