PUBLICATIONS

论文成果

精选近年代表成果；会议版与期刊扩展版分别列出。

完整论文目录请查阅 Google Scholar ↗。论文题名、年份与作者依原文或作者资料核对。

2026

FGCS2026算力与调度

Concord: A GPU cluster scheduler with enhanced interference profiling and asymmetric job packing

Xinhua Wang, Weiwei Lin, Haijie Wu, Jidong Zhai, Keqin Li

通过干扰建模与非对称作业组合，优化 GPU 集群中的资源共享与调度。正式卷期为 2026 年 10 月。

Paper Code

AAAI2026大模型

Experiential Fairness: Bridging the Gap Between User Experience and Resource-Centric Fairness in Online LLM Services

Jiahua Huang, Wentai Wu, Yongheng Liu, Guozhi Liu, Yang Wang, Weiwei Lin

从用户体验出发重新定义大模型服务公平性，并设计轻量级推理调度框架。

Paper Code

IEEE TC2026绿色计算

NoSPF: Non-Stationary Long-Term Power Consumption Forecasting for Servers in Cloud Data Centers

Ruichao Mo, Weiwei Lin, Shengsheng Lin, Simon Fong, Keqin Li

以分层优化刻画功耗序列中的平稳与非平稳成分，研究服务器长期功耗预测。

Paper

IEEE TPAMI2026时间序列

SparseTSF: Lightweight and Robust Time Series Forecasting via Sparse Modeling

Shengsheng Lin, Weiwei Lin, Wentai Wu, Haojun Chen, C. L. Philip Chen

在稀疏时序建模框架中进一步研究轻量化与鲁棒性；为 ICML 2024 工作的期刊扩展版。

Paper Code

AAAI · Poster2026大模型

Are We on the Right Way to Assess Document Retrieval-Augmented Generation?

Wenxuan Shen, Mingjia Wang, Yaochen Wang, Dongping Chen, Junjie Yang, Yao Wan, Weiwei Lin

围绕文档检索增强生成构建评测基准，检查文档 RAG 系统的评估方式。

Paper Code

IEEE IoT-J2026时间序列

SegRNN: Segment Recurrent Neural Network for Long-Term Time Series Forecasting

Shengsheng Lin, Weiwei Lin, Wentai Wu, Feiyu Zhao, Ruichao Mo, Haotong Zhang

通过分段迭代与并行多步预测，构建高效的循环时序预测网络。

Paper Code

IEEE TSC2026边缘智能

PAWSSP: A Two-Stage Parallelism-Aware Algorithm for Joint Workflow Scheduling and Service Placement in Edge Computing

Guanyu Chen, Weiwei Lin, Fang Shi, Haotong Zhang, Bin Wang

利用任务并行性联合优化边缘服务放置与工作流调度，减少应用完成时间。

Paper

Knowledge-Based Systems2026边缘智能

Prodigal: Backdoor defense for federated learning beyond robust aggregation

Guozhi Liu, Weiwei Lin, Tiansheng Huang, Fang Shi, Xiumin Wang, Li Shen

研究联邦学习后门防御中的污染转移现象，通过参数共识与个性化分发提高防御稳健性。

Paper Code

ICML2026大模型

Surgery: Mitigating Harmful Fine-Tuning for Large Language Models via Attention Sink

Guozhi Liu, Weiwei Lin, Tiansheng Huang, Ruichao Mo, Qi Mu, Xiumin Wang, Li Shen

通过注意力沉降机制分析有害微调，并利用正则化抑制大模型学习有害模式。

Paper Code

2025

ICML2025时间序列

Temporal Query Network for Efficient Multivariate Time Series Forecasting

Shengsheng Lin, Haojun Chen, Haijie Wu, Chunyun Qiu, Weiwei Lin

以时间查询捕获全局周期模式，学习多元时间序列中的依赖关系。

Paper Code

IEEE TPDS2025算力与调度

Kairos: Deterministic Scheduling Enhanced by User Collaboration for Deep Learning Workloads

Xinhua Wang, Weiwei Lin, Ruichao Mo, Guozhi Liu, Haijie Wu, Shengjun Tang

结合用户协作与抢占式调度，改善深度学习作业排队时间的可预测性。

Paper Code

IEEE TNNLS2025边缘智能

AdaptiveFL: Communication-Adaptive Federated Learning Under Dynamic Bandwidth

Guozhi Liu, Weiwei Lin, Tiansheng Huang, Fang Shi, Wentai Wu, Li Shen

面向动态带宽环境研究通信自适应联邦学习。

Paper Code

IEEE TSC2025边缘智能

Prediction of Heterogeneous Device Task Runtime Based on Edge Server-Oriented Deep Neuro-Fuzzy System

Haijie Wu, Weiwei Lin, Wangbo Shen, Xiumin Wang, C. L. Philip Chen, Keqin Li

面向边缘服务器预测异构设备任务运行时间，为协同调度提供支持。

Paper

IEEE TSC2025时间序列

MSCNet: Multi-Scale Network with Convolutions for Long-term Cloud Workload Prediction

Feiyu Zhao, Weiwei Lin, Shengsheng Lin, Shaomin Tang, Keqin Li

融合多尺度建模与卷积，研究长时间跨度的云工作负载预测。

Paper Code

IEEE TIFS2025大模型

Targeted Vaccine: Safety Alignment for Large Language Models Against Harmful Fine-Tuning via Layer-Wise Perturbation

Guozhi Liu, Weiwei Lin, Qi Mu, Tiansheng Huang, Ruichao Mo, Yuren Tao, Li Shen

通过针对安全关键层的扰动，提高大模型面对有害微调时的安全性与资源效率。

Paper Code

ACM Computing Surveys2025算力与调度

Public Datasets for Cloud Computing: A Comprehensive Survey

Guozhi Liu, Weiwei Lin, Haotong Zhang, Jianpeng Lin, Shaoliang Peng, Keqin Li

系统梳理云计算公开数据集，为资源管理、调度与系统研究提供数据参考。

Paper Code

IEEE TSUSC2025绿色计算

Power Management Optimization for Data Centers: A Power Supply Perspective

Huikang Huang, Weiwei Lin, Jianpeng Lin, Keqin Li

从供电系统结构与功率模型出发，系统梳理数据中心成本感知的电力管理与绿色能源优化。

Paper

2024

NeurIPS · Spotlight2024时间序列

CycleNet: Enhancing Time Series Forecasting through Modeling Periodic Patterns

Shengsheng Lin, Weiwei Lin, Xinyi Hu, Wentai Wu, Ruichao Mo, Haocheng Zhong

显式建模周期模式，通过周期与残差分解增强时序预测。

Paper Code

FGCS2024绿色计算

Energy-aware virtual machine placement based on a holistic thermal model for cloud data centers

Jianpeng Lin, Weiwei Lin, Wentai Wu, Wenjun Lin, Keqin Li

结合整体热模型研究虚拟机放置，联合考虑计算与制冷能耗。

Paper Code

ICML · Oral2024时间序列

SparseTSF: Modeling Long-term Time Series Forecasting with 1k Parameters

Shengsheng Lin, Weiwei Lin, Wentai Wu, Haojun Chen, Junjie Yang

通过跨周期稀疏预测分离周期性与趋势，以极少参数完成长时间序列预测。

Paper Code