12 2022 档案

【五期邹昱夫】EMNLP（EMNLP'17）Tensor Fusion Network for Multimodal Sentiment Analysis
摘要：EMNLP2017. 本文认为多模态情感分析存在两个挑战。模态间的动态变化：声音、文本、图像之间会相互作用相互影响。模态内的动态变化：对口语化的文本文件进行情感分析非常困难。 因此，为了应对这两个挑战，作者提出了一种新的模型Tensor Fusion Network（张量融合网络，TFN），TFN能 阅读全文
posted @ 2022-12-16 21:28 方班隐私保护小组 阅读(168) 评论(0) 推荐(0) 编辑
【五期邵润东】CCF-A(S&P'22)SHADEWATCHER: Recommendation-guided Cyber Threat Analysis using System Audit Records
摘要：Zeng, Jun, et al. "SHADEWATCHER: Recommendation-guided Cyber Threat Analysis using System Audit Records." 2022 IEEE Symposium on Security and Privacy 阅读全文
posted @ 2022-12-16 17:40 方班隐私保护小组 阅读(265) 评论(0) 推荐(0) 编辑
【五期李伟平】CCF-A(CSS'21) Secure Multi-Party Computation of Differentially Private Heavy Hitters
摘要：Jonas Böhler and Florian Kerschbaum. 2021. Secure Multi-party Computation of Differentially Private Heavy Hitters. In Proceedings ofthe 2021 ACM SIGSA 阅读全文
posted @ 2022-12-16 17:06 方班隐私保护小组 阅读(111) 评论(0) 推荐(0) 编辑
【五期杨志】CCF-A（NeurIPS'21）VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text
摘要：Akbari, Hassan, et al. "Vatt: Transformers for multimodal self-supervised learning from raw video, audio and text." Advances in Neural Information Pro 阅读全文
posted @ 2022-12-16 16:46 方班隐私保护小组 阅读(86) 评论(0) 推荐(0) 编辑
【五期伍捷威】CCF-A(PMLR'21)Clustered Sampling: Low-Variance and Improved Representativity for Clients Selection in Federated Learning
摘要：Fraboni, Yann, et al. "Clustered sampling: Low-variance and improved representativity for clients selection in federated learning." International Conf 阅读全文
posted @ 2022-12-16 14:56 方班隐私保护小组 阅读(516) 评论(0) 推荐(0) 编辑
【五期邵润东】CCF-B(RAID'20)The Limitations of Federated Learning in Sybil Settings
摘要：Fung, Clement, Chris JM Yoon, and Ivan Beschastnikh. "The limitations of federated learning in sybil settings." 23rd International Symposium on Resear 阅读全文
posted @ 2022-12-04 01:37 方班隐私保护小组 阅读(190) 评论(0) 推荐(0) 编辑
【五期李伟平】CCF-C(Neurocomputing'21)An efficient approach for privacy preserving decentralized deep learning models based on secure multi-party computation
摘要：Anh-Tu Tran, The-Dung Luong, Jessada Karnjana, Van-Nam Huynh. An efficient approach for privacy preserving decentralized deep learning models based on 阅读全文
posted @ 2022-12-03 22:59 方班隐私保护小组 阅读(89) 评论(0) 推荐(0) 编辑
【五期邹昱夫】USENIX Security（USENIX Security'21）Systematic Evaluation of Privacy Risks of Machine Learning Models
摘要：The 30th USENIX Security Symposium. 本文贡献主要有两部分，一是提出一种基于修改预测熵的新成员推理攻击方法。二是定义了一个称为隐私风险分数的指标，用来估计每个样本成为训练集中成员的可能性，提供更加细粒度的隐私风险分析。 首先是基于预测熵的成员推理攻击，模型的整个训练 阅读全文
posted @ 2022-12-03 21:46 方班隐私保护小组 阅读(93) 评论(0) 推荐(0) 编辑

点击右上角即可分享
微信分享提示