先进计算与体系结构 先进计算与体系结构
  • 首页
  • 科研工作
  • 科研资源
  • 科研团队
  • 科研成果
  • 学术活动
  • 新闻动态
  • 加入我们
  • 联系我们
ESC

↑↓ 导航 选择
由 Hugo Blox 提供支持
  • 近期学术活动
    • Cornstarch:面向多模态大模型的高效分布式训练系统
    • CDDFuse:面向多模态图像融合的相关性驱动双分支特征分解
    • 面向结构感知的多模态知识图谱补全:TSAM 模型研究
    • APKGC:带注意力惩罚的噪声加强多模态知识图谱补全
    • TEAL:无训练激活稀疏加速大模型推理
    • D²Quant:面向大语言模型的低比特权重量化方法
    • QPO:基于量化与流水线加速的显存高效 DNN 训练
    • DLP:面向大语言模型的动态层级剪枝
    • TD-PTQ:训练动力学对训练后量化鲁棒性的影响
    • L-GreCo:面向高效精准深度学习的逐层自适应梯度压缩
    • 基于 LSTM 网络的不锈钢薄板激光搭接焊熔深光学相干监测研究
    • WLB-LLM:面向大语言模型训练的负载均衡 4D 并行方法
    • QEP:量化误差传播 —— 重新审视逐层训练后量化
    • Cuckoo:面向异构 GPU 深度学习训练的截止时间感知作业打包
    • 大模型驱动的均衡多模态知识图谱补全:LBMKGC 模型研究
    • AGoQ:面向大模型训练的内存高效激活-梯度-优化器量化框架
    • 多类别无监督异常检测中的极简主义设计:Dinomaly 模型研究
    • DIP:面向大模型多模态训练的动态交错流水线系统
    • IGU-LoRA:基于梯度积分与不确定性感知评分的自适应秩分配方法
    • GLVQ:面向低比特大语言模型压缩的分组格型矢量量化器
    • 面向深度学习集群的资源异构感知与利用率增强调度
    • SpinQuant:基于可学习旋转的大语言模型量化
    • 多维感知与智能决策:从单一视觉盲区到多传感器时空孪生
    • R2Q:基于残差精细量化的鲁棒 2 比特大语言模型
    • 基于敏感度指导的训练后量化方法
    • L2-LoRA:通过层特异性正则化改进低秩适配
    • Alpa:面向分布式深度学习的算子间与算子内并行自动化
    • ITQ3_S:高保真 3-bit 大语言模型推理
    • LLM-Streamline:精简冗余层以压缩大语言模型
    • FreeAct:释放大语言模型量化中的激活约束
    • UNISCHED:面向异构用户需求的深度学习训练作业统一调度器
    • MemMat-YOLOX:一种面向DR视频图像的焊缝缺陷智能检测方法
    • SparseLoCo:面向大模型预训练的稀疏化通信高效分布式训练
  • 科研成果
    • When Reinforcement Learning Meets Zero-Determinant: Social Welfare Control in Iterated Game Under Spatial Crowdsourcing
    • xjb: Fast Float to String Algorithm
    • index
    • index
    • index
    • index
    • index
    • index
    • index
    • index
    • index
    • index
    • Low-bit and Sparsified Gradient Communication for Accelerating Distributed Deep Learning with Convergence Guarantees
    • SelCo: Efficient Distributed Multimodal LLM Training with Selective Co-location
    • Social Welfare Control and Incentive for Free-Riding in Cross-Silo FL-based Caching
    • 知识表示增强的多模态知识图谱补全方法
    • An Online Heterogeneous GPU Cluster Scheduler for Multi-Objective Optimization
    • Automated Chain-of-Thought Data Generation for Meteorological Reasoning: A Study Based on WeatherQA
    • Unilateral Control for Social Welfare of Iterated Game in Vehicular Crowdsensing
    • Enhancing Personalized Trip Recommendations with Attractive Route Analysis and Graph Attention Auto-Encoder
    • 基于敏感度指导的训练后量化方法
    • Near-Surface Air Temperature Estimation Based on an Improved Conditional Generative Adversarial Network
    • Sparse Gradient Communication with AlltoAll for Accelerating Distributed Deep Learning
    • Compact Estimator for Streaming Triangle Counting
    • Implementation of script system based on recording and playback with EventBus in industrial software
    • Design and implementation of script module with recording and playback function in CAE software
    • A Distributed Streaming Framework for Edge-Cloud Triangle Counting
    • Realizing Power-efficient Encryption Communication for Computational RFID Tags
    • ACM: Accuracy-aware Collaborative Monitoring for Software-Defined Network-wide Measurement
    • 基于Mask R-CNN卷积神经网络的虹膜分割
    • Multi-Scale Progressive Blind Face Deblurring
    • Pedestrian Flow Prediction and Route Recommendation with Business Events
    • Pluralistic Face Inpainting with Transformation of Attribute Information
    • Generation Rate Control with AoI under Traffic Hole Problem in Vehicular Networks
    • Make U-Net Greater: An Easy-to-Embed Approach to Improve Segmentation Performance Using Hypergraph
    • A Multi-Task Network for Cardiac Magnetic Resonance Image Segmentation and Classification
    • Distributed Triangle Approximately Counting Algorithms in Simple Graph Stream
    • A novel offloading strategy for high speed mobile and data intensive intelligent sensor
    • Enhancing Personalized Trip Recommendation with Attractive Routes
    • Parallelization and I/O Performance Optimization of a Global Nonhydrostatic Dynamical Core using MPI
    • Distributed Triangle Counting Algorithms in Simple Graph Stream
    • Pedestrian Flow Prediction with Business Events
    • Towards Cascading Problem for Dynamic Rate Allocations in ISP Networks with SDN
    • Optimization of parallel program based on lattice BGK method
    • A Heterogeneous Virtual Machines Resource Allocation Scheme in Slices Architecture of 5G Edge Datacenter
    • Molecular Dynamics Simulation Optimization Based on GROMACS on Sunway TaihuLight
    • A Parallel Optimization of the Fast Algorithm of Convolution Neural Network on CPU
    • Crowdsourcing-Based Framework for Teaching Quality Evaluation and Feedback Using Linguistic 2-Tuple
    • Deployment Mechanism Design for Cost-Effective Data Uploading in Delay-Tolerant Crowdsensing
    • Efficient Complex Event Processing under Boolean Model
    • Query privacy in pervasive computing
    • New privacy protect algorithm for pervasive computing
    • Overview of location privacy in pervasive computing
    • Research on search radius limited location algorithm in distributed spanning tree
    • Fault tolerance on improved distributed spanning tree structure
    • Improved Distributed Spanning Tree and Case Study for Replica Location Service
    • Location privacy in sparse environment
    • 一个普适计算环境中的定位隐私保护模型
    • 基于分布式生成树的副本定位服务
  • 博客
    • 基于Spark引擎和FNLP的微博用户行为大数据分析系统介绍
  • 研究方向
    • 大模型压缩与高效推理
    • 大数据挖掘与智能计算
    • 分布式机器学习与系统优化
    • 智能图像处理与计算机视觉
  • 科研资源与基础设施
  • 科研项目
    • 数据编织分布式存储关键技术研究
    • 国产高性能算力芯片研发
  • 加入我们

2026年1月1日·
梁浩轩
梁浩轩
· 0 分钟阅读时长
类型
期刊文章
publications
最近更新于 2026年1月1日
Sina Weibo
梁浩轩
Authors
梁浩轩 (he/him)
硕士研究生
2024年9月进入成都信息工程大学读研,发表学术论文一篇。

← 2026年1月1日
2026年1月1日 →
语言:
中文 (简体)
English

© 2026 先进计算与智能信息处理实验室 蜀ICP备2026005197号-1