按年被提到
20172018 · 1 次2026
AI 安全、伦理与社会影响›模型窃取、记忆与水印·arXiv:1710.06963 · cs.LG

Learning Differentially Private Recurrent Language Models

H. Brendan McMahan, Daniel Ramage, Kunal Talwar, Li Zhang · 2017
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2018-03-25 · 最近 2019-05-18
2不同的人
2活跃月份
2讨论串
849被引(OpenAlex)
arXiv 摘要(候选层不写导读)

We demonstrate that it is possible to train large recurrent language models with user-level differential privacy guarantees with only a negligible cost in predictive accuracy. Our work builds on recent advances in the training of deep networks on user-partitioned data and privacy accounting for stochastic gradient descent. In particular, we add user-level privacy protection to the federated averaging algorithm, which makes "large step" updates from user-level data. Our work demonstrates that given a dataset with a sufficiently large number of users (a requirement easily met by even small internet-scale datasets), achieving differential privacy comes at the cost of increased computation, rather than in decreased utility as in most prior work. We find that our private LSTM language models are quantitatively and qualitatively similar to un-noised models when trained on a large dataset.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
评论singhrac · 2018-03

“One problem I've found with differential privacy is that no one talks about how to set \epsilon. I've read this book, and it's quite well written and complete, but as the title says it focuses on the algorithmic foundations. This paper [1] is much better for practitioners, and actually gives very reasonable values for the privacy guarantee (e.g., (1.2, 1…”

The Algorithmic Foundations of Differential Privacy (2014) [pdf]原评论 ↗
评论CyanTas · 2019-05

“You’re not wrong, but it does say those privacy risks can be mitigated with differential privacy. That McMahan et al. paper (which is also Google) makes the accuracy cost seem low. https://arxiv.org/abs/1710.06963”

Federated Learning原评论 ↗