arXivDaily arXiv每日学术速递 周一至周五更新

大厂专区

Google(谷歌)

2026-07-13 至 2026-07-13 共收录 2
2606.03979 2026-07-13 cs.LG cs.AI 版本更新

Language Models Need Sleep: Learning to Self-Modify and Consolidate Memories

语言模型需要睡眠:学习自我修改和巩固记忆

Ali Behrouz, Farnoosh Hashemi, Adel Javanmard, Vahab Mirrokni

机构 * Google(谷歌) Cornell University(康奈尔大学)

AI总结 受人类学习过程启发,提出“睡眠”范式,通过记忆巩固(知识播种)和梦境(自我改进)两阶段,使模型持续学习、将短期记忆转化为长期知识并自我提升。

Comments A version of this work has been publicly available from September 2025 on OpenReview

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.03238 2026-07-13 cs.CR cs.AI cs.LG stat.ML 版本更新

How to DP-fy Your Data: A Practical Guide to Generating Synthetic Data With Differential Privacy

如何让你的数据实现差分隐私:使用差分隐私生成合成数据的实用指南

Natalia Ponomareva, Zheng Xu, H. Brendan McMahan, Peter Kairouz, Lucas Rosenblatt, Vincent Cohen-Addad, Cristóbal Guzmán, Ryan McKenna, Galen Andrew, Alex Bie, Da Yu, Alex Kurakin, Morteza Zadimoghaddam, Sergei Vassilvitskii, Andreas Terzis

机构 * Google Research USA(谷歌DeepMind) NYU Work done at Google as part of student researcher engagement New York NY USA Google Research New York NY USA Institute for Mathematical Computational Engineering, Faculty of Mathematics School of Engineering, Pontificia Universidad Cat\'olica de Chile Santiago Chile Google DeepMind Mountain View CA USA Google Research School of Engineering, Pontificia Universidad Cat\'olica de Chile Google DeepMind

AI总结 研究如何利用差分隐私生成合成数据,探讨相关技术、保护类型及不同模态进展,概述系统组件,旨在推动差分隐私合成数据应用,促进研究并增强信任。

Journal ref JAIR 2026, vol. 86 JAIR, Vol. 86

详情

展开后加载摘要…

URL PDF HTML 收藏