会议 | 李萌的个人主页

一篇论文被IEEE S&P (Oakland)接收

Tue, 15 Sep 2026 00:00:00 +0800

论文题目为“Cachemir: Fully Homomorphic Encrypted Inference of Generative Large Language Model with KV Cache”。该工作由Ye Yu主导。

一篇论文被TCAD'2026接收，一篇论文被CoRL'2026接收

Sat, 05 Sep 2026 00:00:00 +0000

被TCAD接收的论文题目为“HDA-MoE: Hybrid Parallelism and Dynamic, Adaptive Scheduling for Mixture-of-Experts with 3D Near-Memory Processing”，提出了一种面向3D DRAM加速器上MoE推理的新型调度算法。该工作是ICCAD'2025论文HD-MoE的期刊扩展版本，由博士生Haochen Huang主导。

被CoRL接收的论文题目为“Jetson-PI: Towards Onboard Real-Time Robot Control via Foresight-Aligned Asynchronous Inference”，提出了异步推理优化和边缘GPU部署优化方法。该工作由博士生Zebin Yang主导。

一篇论文被CCS'2026接收

Fri, 17 Jul 2026 00:00:00 +0000

论文题目为“ROSETTA: Efficient and Accurate Privacy-Preserving LLM Decoding via Hybrid CKKS/TFHE Evaluation”，提出了一种通过混合CKKS/TFHE方案计算大语言模型非线性函数的新方法。该工作由博士生Jiangrui Yu主导。

两篇论文被MICRO 2026接收，一篇论文被ICCAD 2026接收

Sat, 11 Jul 2026 00:00:00 +0000

被接收的论文如下：

OptiPrime: Optimizing Private Inference through Protocol-Hardware Codesign（由Jiangrui Yu主导，与来自OpenSecurity Research的Si Chen和Junfeng Fan合作）
Helios: Melting Kernel Boundaries for GPU-Accelerated HE via Graph Rewriting and Microarchitecture-Aware Mapping（由Yi Chen和Ziyu Tang主导，与来自Ant Group的Mingzhe Zhang合作）
MatMoE: Matryoshka Mixture-of-Experts with Dynamic Mixed-Precision Quantization for Efficient Inference（由Renjie Wei主导）

三篇论文被ICML 2026接收

Fri, 01 May 2026 00:00:00 +0000

被接收的论文如下：

Attention Sink Forges Native MoE in Attention Layers: Sink-Aware Training to Address Head Collapse（由Zizhuo Fu主导）
TEAM: Temporal–Spatial Consistency Guided Expert Activation for MoE Diffusion Language Model Acceleration（由Linye Wei主导）
HyPER: Bridging Exploration and Exploitation for Scalable LLM Reasoning with Hypothesis Path Expansion and Reduction（由Shengxuan Qiu、Haochen Huang主导）

一篇论文被OSDI'2026接收，两篇合作论文被ISCA'2026接收

Mon, 30 Mar 2026 00:00:00 +0000

OSDI'2026论文：

Breaking the Reward Barrier: Accelerating Tree-of-Thought Reasoning via Speculative Exploration

ISCA'2026论文：

DIAMoND: Dynamic Inference for Adaptive Edge MoE with Heterogeneous In-NAND and Near-DRAM Compute Architecture（由Prof. Ling Liang主导）
PipeComm: Maximizing Link Utilization through Pipeline-Aware Collective Communication Synthesis（由Prof. Yun Liang主导）

七篇论文被DAC'2026接收

Tue, 24 Feb 2026 00:00:00 +0000

高效人工智能加速器：

EdgeSC: Universal Stochastic Computing Architecture for Efficient Edge Detection（与来自HKU的Prof. Ngai Wong合作）
DRIFT: Harnessing Inherent Fault Tolerance for Efficient and Reliable Diffusion Model Inference
NASiC: 3D NAND-based CAM-Selected Multibit CIM Architecture for Efficient On-Device Mixture-of-Experts LLM Inference
S2CIM: A Secure-Computation and Secure-Storage Compute-in-Memory Architecture with Circuit-Algorithm Co-Design for Efficient and Trustworthy Edge Inference（由来自PKU的Prof. Kechao Tang主导）

高效人工智能算法：

KEEP: A KV-Cache-Centric Memory Management System for Efficient Embodied Planning
Orchestrating Dual-Boundaries: An Arithmetic Intensity Inspired Acceleration Framework for Diffusion Language Models
DySL-VLA: Efficient Vision-Language-Action Model Inference via Dynamic-Static Layer-Skipping for Robot Manipulation

Zizhuo Fu获得ACM/SIGDA学生科研竞赛第一名

Sun, 02 Nov 2025 00:00:00 +0000

一篇论文被ASPLOS'2026接收

Sat, 01 Nov 2025 00:00:00 +0000

一篇论文被ASSCC'2025接收

Tue, 30 Sep 2025 00:00:00 +0000

三篇论文被NeurIPS 2025接收

Thu, 18 Sep 2025 00:00:00 +0000

一篇PPML综述论文发布于arXiv

Sun, 10 Aug 2025 00:00:00 +0000

Zizhuo Fu获得商汤奖学金

Fri, 01 Aug 2025 00:00:00 +0000

两篇论文分别被MICRO'2025和TACO'2025接收

Thu, 17 Jul 2025 00:00:00 +0000

五篇论文和一个专题研讨被ICCAD'2025接收

Wed, 02 Jul 2025 00:00:00 +0000

高效人工智能方向论文：

No Redundancy, No Stall: Lightweight Streaming 3D Gaussian Splatting for Real-time Rendering
H2EAL: Hybrid-Bonding Architecture with Hybrid Sparse Attention for Efficient Long-Context LLM Inference
HD-MoE: Hybrid and Dynamic Parallelism for Mixture-of-Expert LLMs with 3D Near-Memory Processing
SpecMamba: Accelerating Mamba Inference on FPGA with Speculative Decoding
Adder-DCIM: A Parallel Bit-Flexible Digital CIM Joint Model Compression Framework for AdderNet Inference（由Prof. Xiyuan Tang主导）

一篇论文被USENIX Security'2025接收

Thu, 15 May 2025 00:00:00 +0000

一篇合作论文被IEEE TIFS'2025接收

Sun, 02 Mar 2025 00:00:00 +0000

四篇论文被DAC'2025接收

Sat, 15 Feb 2025 00:00:00 +0000

四篇关于高效大语言模型的论文：

HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference
SpecASR: Accelerating LLM-based Automatic Speech Recognition via Speculative Decoding
UniCAIM: A Unified CAM/CIM Architecture with Static-Dynamic KV Cache Pruning for Efficient Long-Context LLM Inference
ReaLM: Reliable and Efficient Large Language Model Inference with Statistical Algorithm-Based Fault Tolerance

四篇论文被DATE'2025接收

Sat, 09 Nov 2024 00:00:00 +0000

隐私保护人工智能方向论文：

FLASH: An Efficient Hardware Accelerator Leveraging Approximate and Sparse FFT for Homomorphic Encryption

高效人工智能方向论文：

LightMamba: Efficient Mamba Acceleration on FPGA with Quantization and Hardware Co-design
SCALES: Boost Binary Neural Network for Image Super-Resolution with Efficient Scalings
Compact Non-Volatile Lookup Table Architecture based on Ferroelectric FET Array through In-Situ Combinatorial One-Hot Encoding for Reconfigurable Computing

两篇论文被NeurIPS 2024接收

Thu, 26 Sep 2024 00:00:00 +0000

一篇论文被IEEE TCASI'2024接收

Tue, 03 Sep 2024 00:00:00 +0000

李萌博士在CCF Chip 2024作三场报告

Sat, 10 Aug 2024 00:00:00 +0000

李萌博士在CCF Chip 2024围绕高效与隐私保护人工智能作了三场报告，包括：

通过网络/协议协同优化实现高效、隐私保护的Transformer推理
面向高效大语言模型推理的编译器优化
高效且准确的端到端随机计算人工智能加速

一篇论文被MICRO 2024接收

Sat, 10 Aug 2024 00:00:00 +0000

七篇论文被ICCAD'2024接收

Thu, 27 Jun 2024 00:00:00 +0000

隐私保护人工智能方向论文：

PrivQuant: Communication-Efficient Private Inference with Quantized Network/Protocol Co-Optimization
FlexHE: A flexible Kernel Generation Framework for Homomorphic Encryption-Based Private Inference

高效人工智能方向论文：

OSCA: End-to-end Serial Stochastic Computing Neural Acceleration with Fine-grained Scaling and Piecewise Activation
HG-PIPE: Vision Transformer Acceleration with Hybrid-Grained Pipeline
ProPD: Dynamic Token Tree Pruning and Generation for LLM Parallel Decoding
AdapMoE: Adaptive Sensitivity-based Expert Gating and Management for Efficient MoE Inference
MCUBERT: Memory-Efficient BERT Inference on Commodity Microcontrollers

三篇论文被DAC'2024接收

Fri, 16 Feb 2024 00:00:00 +0000

一篇论文被IEEE CICC'2024接收

Mon, 08 Jan 2024 00:00:00 +0000

一篇论文被IEEE TCASI'2024接收

Mon, 08 Jan 2024 00:00:00 +0000

一篇论文被DATE 2024接收

Thu, 09 Nov 2023 00:00:00 +0000

一篇论文被3DV 2024接收

Fri, 20 Oct 2023 00:00:00 +0000

一篇论文被NeurIPS 2023接收

Sat, 23 Sep 2023 00:00:00 +0000

三篇论文被ICCAD 2023接收

Mon, 24 Jul 2023 00:00:00 +0000

一篇论文被ICCV 2023接收

Mon, 24 Jul 2023 00:00:00 +0000

担任DAC'23和GLSVLSI'23程序委员会委员及DAC'23 Student Design Contest and Ph.D. Forum共同主席

Tue, 07 Mar 2023 00:00:00 +0000

一篇论文被IEEE TCAD接收

Tue, 07 Mar 2023 00:00:00 +0000

一篇论文被DAC 2023接收

Wed, 15 Feb 2023 00:00:00 +0000

两篇论文被DATE 2023接收

Mon, 21 Nov 2022 00:00:00 +0000

一篇论文被NeurIPS 2022接收

Thu, 15 Sep 2022 00:00:00 +0000

一篇论文被ICML 2022接收

Wed, 18 May 2022 00:00:00 +0000

一篇论文被IJCAI-ECAI 2022接收

Sun, 15 May 2022 00:00:00 +0000

两篇论文被CVPR 2022接收

Sat, 05 Mar 2022 00:00:00 +0000

一篇论文被ICASSP 2022接收

Tue, 01 Mar 2022 00:00:00 +0000