Google TechTalks
Google TechTalks — 292 videos with subtitlesGoogle TechTalks has 292 videos indexed on WinkLingo (channel with 349K subscribers). Every video is searchable word by word.
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…
Recent videos
58:08Local Node Differential Privacy
57:54Privacy Amplification for Correlated-Noise Mechanisms via b-Min-Sep Subsampling
40:17Differentially Private Table-Image Multimodal Data Generation
56:22Atomic Facts to Structured Knowledge: Rethinking Unlearning & Jailbreaking in Large Language Models
18:32Understanding Trade-offs Between Worst-case Differential Privacy Guarantees & Real Threat Models
53:30Machine Text Detectors are Membership Inference Attacks
57:57Code Health Guardian
47:17Going Back and Beyond: Emerging (Old) Threats in LLM Privacy and Poisoning
33:40POPri: Private Federated Learning using Preference-Optimized Synthetic Data
33:23Leveraging Per-Instance Privacy for Machine Unlearning
50:30Private Adaptations of Large Language Models
40:16Disparate Privacy Risks from Medical AI - An Investigation into Patient-level Privacy Risk
54:48Stable Estimators for Fast Private Statistics
57:39Worst-Case Membership Inference of Language Models
38:13Cascading Adversarial Bias from Injection to Distillation in Language Models
45:47The Limits and Possibilities of One Run Auditing
33:10Persistent Pre-Training Poisoning of LLMs
53:45Privacy Auditing of Large Language Models
32:31Cascading Adversarial Bias from Injection to Distillation in Language Models
45:45Differentially Private Synthetic Data without Training
49:31Threat Models for Memorization: Privacy, Copyright, and Everything In-Between
42:03Privacy Ripple Effects from Adding or Removing Personal Information in Language Model Training
51:48Streaming Private Continual Counting via Binning
51:27Chasing the Constants and its Implications in Differential Privacy