六月 19, 2025
实验室有5篇论文被EDA领域旗舰会议ICCAD接收。所有论文均围绕高效AI算法和芯片展开:
- No Redundancy, No Stall: Lightweight Streaming 3D Gaussian Splatting for Real-time Rendering (一作为大四本科生韦临烨)
- H2EAL: Hybrid-Bonding Architecture with Hybrid Sparse Attention for Efficient Long-Context LLM Inference(一作为大三本科生傅子酌、二年级博士生郭笑天)
- HD-MoE: Hybrid and Dynamic Parallelism for Mixture-of-Expert LLMs with 3D Near-Memory Processing(一作为大三本科生黄浩宸、二年级博士生仲书璋)
- SpecMamba: Accelerating Mamba Inference on FPGA with Speculative Decoding(一作为二年级硕士生钟林峰、二年级硕士生许颂强)
- Adder-DCIM: A Parallel Bit-Flexible Digital CIM Joint Model Compression Framework for AdderNet Inference (通讯作者为Prof. Xiyuan Tang)
同时,实验室还首次在ICCAD中组织Special Session,欢迎关注。