Unverified50% confidenceFactExact time
Differential privacy is a rigorous mathematical definition of privacy: a randomized algorithm satisfies (ε, δ)-DP if the difference in output distributions on any adjacent datasets is constrained by ε and δ; the smaller ε is, the stronger the privacy protection, but usually at the cost of degraded model utility.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedDifferential Privacy was formally defined by Cynthia Dwork in 2006, with its core guarantee being that the output of an algorithm will not change significantly due to the addition or removal of any single record in the dataset.79% similarUnverifiedDifferentially private stochastic gradient descent (DP-SGD) has become the de facto standard for training privacy-sensitive models.72% similarUnverified差分隐私由Cynthia Dwork等人于2006年提出，其核心定义为P[M(D)∈S] ≤ e^ε × P[M(D')∈S]71% similarVerifiedIn differential privacy, a smaller privacy budget ε means stronger privacy protection but also a corresponding decrease in data utility.71% similarUnverifiedDifferential privacy is currently the most mainstream privacy protection mechanism, providing mathematically provable privacy guarantees by injecting calibrated noise, and has already been deployed at scale by companies such as Apple and Google.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1020472API
curl https://kongchang.com/api/v1/knowledge/claims/1020472MCP
get_claim(id=1020472)