SJTU AISIG

SJTU Artificial Intelligence Special Interest Group

研究方向

SJTU AISIG 团队围绕分布式计算与无服务器计算为主要研究方向，致力于构建智能云计算系统。

围绕分布式计算展开联邦学习和并行算法两方面的研究。联邦学习方面研究系统安全性、隐私性、异构性和高效性：安全性主要研究存在拜占庭客户端时，保障系统收敛性；隐私性主要研究如何防御针对模型的隐私攻击；异构性主要研究针对数据异构的个性化联邦学习，和针对设备异构的系统资源调度优化；高效性主要研究模型训练过程中的通信及计算优化。并行算法方面针对于单并行、多并行和自动并行三方面展开，寻找高效的并行方案。

围绕无服务器计算展开应用部署、资源调度和系统安全三方面的研究。应用部署方面研究特定应用在无服务器计算平台的部署和优化，包括机器学习训练与推理、MapReduce大数据处理等。资源调度方面研究如何为无服务器计算提供细粒度的系统资源调度机制，包括空闲资源收集、异构资源等。系统安全研究如何保证轻量级应用的安全执行机制，包括资源隔离、安全监控、认证授权与数据保护等。


新闻速递

2021年11月 论文 Improving Bayesian Neural Networks by Adversarial Sampling 被 AAAI’ 22 接收，恭喜家儒！
2021年08月 论文 Siren: Byzantine-robust Federated Learning via Proactive Alarming 被 SoCC’ 21 接收，恭喜含熙！
2021年04月 论文 Themis: A Fair Evaluation Platform for Computer Vision Competitions 被 IJCAI’ 21 接收，恭喜子诺！
2021年03月 论文 Robust Bayesian Neural Networks by Spectral Expectation Bound Regularization 被 CVPR’ 21 接收，恭喜家儒！

代表工作

  1. SoCC (CCF-B)
    Siren: Byzantine-robust Federated Learning via Proactive Alarming
    Guo Hanxi, Wang Hao, Song Tao, Hua Yang, Lv Zhangcheng, Jin Xiulang, Xue Zhengui, Ma Ruhui, and Guan Haibing
    In Proceedings of the ACM Symposium on Cloud Computing 2021
  2. IJCAI (CCF-A)
    Themis: A Fair Evaluation Platform for Computer Vision Competitions
    Cai Zinuo, Yuan Jianyong, Hua Yang, Song Tao, Wang Hao, Xue Zhengui, Hu Ningxin, Ding Jonathan, Ma Ruhui, Haghighat Mohammad Reza, and Guan Haibing
    In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence 2021
  3. IEEE Network (CCF-A)
    SpaceDML: Enabling Distributed Machine Learning in Space Information Networks
    Guo Hanxi, Yang Qing, Wang Hao, Hua Yang, Song Tao, Ma Ruhui, and Guan Haibing
    IEEE Network 2021
  4. CVPR (CCF-A)
    Robust Bayesian Neural Networks by Spectral Expectation Bound Regularization
    Zhang Jiaru, Hua Yang, Xue Zhengui, Song Tao, Zheng Chengyu, Ma Ruhui, and Guan Haibing
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2021
  5. ICCV (CCF-A)
    Self-Supervised Vessel Segmentation via Adversarial Learning
    Ma Yuxin, Hua Yang, Deng Hanming, Song Tao, Wang Hao, Xue Zhengui, Cao Heng, Ma Ruhui, and Guan Haibing
    In Proceedings of the IEEE/CVF International Conference on Computer Vision 2021
  6. ECCV (CCF-B)
    Dual adversarial network for deep active learning
    Wang Shuo, Li Yuexiang, Ma Kai, Ma Ruhui, Guan Haibing, and Zheng Yefeng
    In European Conference on Computer Vision 2020
  7. AAAI (CCF-A)
    Reinforcing Neural Network Stability with Attractor Dynamics
    Deng Hanming, Hua Yang, Song Tao, Xue Zhengui, Ma Ruhui, Robertson Neil, and Guan Haibing
    In Proceedings of the AAAI Conference on Artificial Intelligence 2020
  8. ACM MM (CCF-A)
    Unsupervised video summarization with attentive conditional generative adversarial networks
    He Xufeng, Hua Yang, Song Tao, Zhang Zongpu, Xue Zhengui, Ma Ruhui, Robertson Neil, and Guan Haibing
    In Proceedings of the 27th ACM International Conference on multimedia 2019