TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Differential Privacy

136 点作者 sr2将近 8 年前

7 条评论

eddyg将近 8 年前
This[0] video from Apple&#x27;s WWDC gives a nice overview of how Differential Privacy is being used in iOS. Basically, Apple can collect and store its users’ data in a format that lets it glean useful info about what people do, say, like and want. But it <i>can&#x27;t</i> extract anything about a single specific one of those people that might represent a privacy violation. And neither can hackers or intelligence agencies.<p>[0] <a href="https:&#x2F;&#x2F;developer.apple.com&#x2F;videos&#x2F;play&#x2F;wwdc2016&#x2F;709&#x2F;?time=812" rel="nofollow">https:&#x2F;&#x2F;developer.apple.com&#x2F;videos&#x2F;play&#x2F;wwdc2016&#x2F;709&#x2F;?time=8...</a> (the &quot;Transcript&quot; tab has the text of the video if you want to read instead of watch.)
评论 #14624442 未加载
JoachimSchipper将近 8 年前
I like <a href="https:&#x2F;&#x2F;blog.cryptographyengineering.com&#x2F;2016&#x2F;06&#x2F;15&#x2F;what-is-differential-privacy&#x2F;" rel="nofollow">https:&#x2F;&#x2F;blog.cryptographyengineering.com&#x2F;2016&#x2F;06&#x2F;15&#x2F;what-is-...</a> as an introduction.<p>Differential privacy is cool. However, I looked at Google&#x27;s RAPPOR algorithm (deployed in Chrome, and clearly designed with real-world considerations in mind) in some depth, and I found that RAPPOR needs millions to billions of measurements to become useful, even while exposing users to potentially serious security risks (epsilon = ln(3), so &quot;bad things become at most 3x more likely&quot;). Much better than doing nothing, but we&#x27;ll continue to need non-cryptographic solutions (NDA&#x27;s etc.) for many cases.
BucketSort将近 8 年前
The coolest part about differential privacy is its guarantees about over fitting.
评论 #14623399 未加载
评论 #14621699 未加载
jey将近 8 年前
I think this is the canonical review article: <a href="https:&#x2F;&#x2F;www.cis.upenn.edu&#x2F;~aaroth&#x2F;Papers&#x2F;privacybook.pdf" rel="nofollow">https:&#x2F;&#x2F;www.cis.upenn.edu&#x2F;~aaroth&#x2F;Papers&#x2F;privacybook.pdf</a><p>(No, I haven&#x27;t read it...)
评论 #14624213 未加载
cjbprime将近 8 年前
I don&#x27;t like differential privacy very much.<p>Take GPS data, for example: NYC has released a taxicab dataset showing the &quot;anonymized&quot; location of every pickup and dropoff.<p>This is bad for privacy. One attack is that now if you know when and where someone got in a cab (perhaps because you were with them when they got in), you can find out if they were telling the truth to you about where they were going -- if there are no hits in the dataset showing a trip from the starting location that you know to the ending location that they claimed, then they didn&#x27;t go where they said they did.<p>Differential privacy researchers claim to help fix these problems by making the data less granular, so that you can&#x27;t unmask specific riders: blurring the datapoints so that each location is at a city block&#x27;s resolution, say. But that doesn&#x27;t help in this case -- if no-one near the starting location you know went to the claimed destination, blurring doesn&#x27;t help to fix the information leak. You didn&#x27;t <i>need</i> to unmask a specific rider to disprove a claim about the destination of a trip.<p>I think that flaws like these mean that we should just say that GPS trip data is &quot;un-de-identifiable&quot;. I suspect the same is true for all sorts of other data. For example, Y chromosomes are inherited the same way that surnames often are, meaning that you can make a good guess at the surname of a given &quot;deidentified&quot; DNA sequence, and thus unmask its owner from a candidate pool, given a genetic ancestry database of the type that companies are rapidly building.
评论 #14621878 未加载
评论 #14620421 未加载
评论 #14621117 未加载
评论 #14621984 未加载
评论 #14620847 未加载
projectramo将近 8 年前
At one point, I know someone who wanted to give money to a large medical organization so that they could show their patients the tradeoff between various interventions. (efficacy vs side-effects).<p>It was going to be donated money to build an app that belonged to the institution.<p>The institution would not let their own researches publish the data on the app even though it was anonymous. They didn&#x27;t want to take the risk.<p>It would be great if this lead to accepted protocols that made it so that people didn&#x27;t have to think about it. &quot;Oh yeah, we&#x27;ll share it using DP&quot; and then people could move ahead using data.
rectang将近 8 年前
Shades of the AOL search data leak:<p><a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;AOL_search_data_leak" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;AOL_search_data_leak</a><p><i>Of course</i> personally identifiable information will be extracted despite this model. &quot;Differential Privacy&quot; is cynical academic malpractice -- selling a reputation so that when individuals are harmed in the course of commercial exploitation of the purportedly anonymized data, the organizations that profited can avoid being held responsible.<p>We never learn, because there is money to be made if we pretend that anonymization works.
评论 #14621327 未加载
评论 #14620895 未加载
评论 #14626689 未加载
评论 #14621322 未加载