Google TechTalks
Google TechTalks · 292 個帶字幕的影片Google TechTalks 在 WinkLingo 收錄了 292 個影片(這個頻道有 35萬 個訂閱),每個影片都能逐字檢索。
查看 YouTube 原頻道英文簡介
Google Tech Talks is a grass-roots program at Google for sharing information of interest to the technical community. At its best, it's part of an ongoing discussion about our world featuring top experts in diverse fields. Presentations range from the broadest of perspective overviews to the most technical of deep dives, on topics well-established to wildly speculative. Disclaimer: Google TechTalks are designed to disseminate a wide spectrum of views on topics including Current Affairs, Science, Medicine, Engineering, Business, Humanities, Law, Entertainment, and the Arts. The views or…
收錄的影片
58:08本地節點差分隱私
Local Node Differential Privacy
57:54通過 b-Min-Sep 子取樣實現相關噪聲機制的隱私增強
Privacy Amplification for Correlated-Noise Mechanisms via b-Min-Sep Subsampling
40:17差分隱私的表格-影像多模態資料生成
Differentially Private Table-Image Multimodal Data Generation
56:22從原子事實到結構化知識:重新思考大語言模型中的“取消學習”與“越獄”
Atomic Facts to Structured Knowledge: Rethinking Unlearning & Jailbreaking in Large Language Models
18:32理解最差情況差分隱私保證與真實威脅模型之間的權衡
Understanding Trade-offs Between Worst-case Differential Privacy Guarantees & Real Threat Models
53:30機器文本檢測器本質上是成員推理攻擊
Machine Text Detectors are Membership Inference Attacks
57:57程式碼健康守護者 (Code Health Guardian)
Code Health Guardian
47:17回顧與前瞻:LLM 隱私與投毒中出現的新(舊)威脅
Going Back and Beyond: Emerging (Old) Threats in LLM Privacy and Poisoning
33:40POPri:使用偏好優化合成資料的私有聯邦學習
POPri: Private Federated Learning using Preference-Optimized Synthetic Data
33:23利用單例項隱私實現機器解除安裝 (Machine Unlearning)
Leveraging Per-Instance Privacy for Machine Unlearning
50:30大語言模型的私有化適配
Private Adaptations of Large Language Models
40:16醫療 AI 帶來的差異化隱私風險——患者級隱私風險調查
Disparate Privacy Risks from Medical AI - An Investigation into Patient-level Privacy Risk
54:48用於快速私有統計的穩定估計量
Stable Estimators for Fast Private Statistics
57:39語言模型的最壞情況成員推理 (Membership Inference)
Worst-Case Membership Inference of Language Models
38:13語言模型中從注入到蒸餾的級聯對抗偏差
Cascading Adversarial Bias from Injection to Distillation in Language Models
45:47單次執行審計(One Run Auditing)的侷限性與可能性
The Limits and Possibilities of One Run Auditing
33:10大語言模型 (LLM) 的持久預訓練中毒問題
Persistent Pre-Training Poisoning of LLMs
53:45大語言模型的隱私審計
Privacy Auditing of Large Language Models
32:31語言模型中從注入到蒸餾的級聯對抗偏差
Cascading Adversarial Bias from Injection to Distillation in Language Models
45:45無需訓練的差分隱私合成資料
Differentially Private Synthetic Data without Training
49:31記憶的威脅模型:隱私、版權及其之間的一切
Threat Models for Memorization: Privacy, Copyright, and Everything In-Between
42:03語言模型訓練中新增或刪除個人資訊帶來的隱私連鎖反應
Privacy Ripple Effects from Adding or Removing Personal Information in Language Model Training
51:48通過分桶實現私有連續計數的流式傳輸
Streaming Private Continual Counting via Binning
51:27追尋常數及其在差分隱私中的影響
Chasing the Constants and its Implications in Differential Privacy