更改

差分隐私 (查看源代码)

2021年11月7日 (日) 04:17的版本

删除59字节、 2021年11月7日 (日) 04:17

小

edition on 06/11/21

第5行：第5行：

Differential privacy is a system for publicly sharing information about a dataset by describing the patterns of groups within the dataset while withholding information about individuals in the dataset. The idea behind differential privacy is that if the effect of making an arbitrary single substitution in the database is small enough, the query result cannot be used to infer much about any single individual, and therefore provides privacy. Another way to describe differential privacy is as a constraint on the algorithms used to publish aggregate information about a statistical database which limits the disclosure of private information of records whose information is in the database. For example, differentially private algorithms are used by some government agencies to publish demographic information or other statistical aggregates while ensuring confidentiality of survey responses, and by companies to collect information about user behavior while controlling what is visible even to internal analysts.

−

是一个用于公开分享数据集信息的系统，它描述了数据集中的群体模式，同时保留了数据集中个人的信息。差分隐私的想法是，如果在数据库中任意进行单一替换的效果足够小，那么查询结果就不能用于推断任何单个个体的太多信息，因此提供了隐私。另一种描述差分隐私数据库的方法是限制用于发布统计数据库的聚合信息的算法，这种算法限制了数据库中信息的记录的私人信息的披露。例如，一些政府机构使用不同的私有算法公布人口统计信息或其他统计数据，同时确保调查答复的保密性，而公司则使用私有算法收集用户行为的信息，同时控制哪怕是内部分析师也能看到的信息。

+

差分隐私是一个用于公开分享数据集信息的系统，它在描述数据集中的群体特征的同时保护了数据集中的个人信息。差分隐私的理念是，如果在数据库中进行任意单次更迭的影响足够小，那么查询结果就不能用于推断任何单一个体的大量信息，因此个体的隐私得以保证。另一种对于差分隐私的描述表示，这是针对发布统计数据库的聚合信息的算法的约束条件，这种算法限制了数据库中信息的记录的个人信息的披露。例如，一些政府机构使用差分隐私算法公布人口信息或其他统计数据，同时确保调查结果的保密性；而公司则使用该算法收集用户行为信息，同时控制哪些信息对于内部分析人员是可见的。

Roughly, an algorithm is differentially private if an observer seeing its output cannot tell if a particular individual's information was used in the computation.

第13行：第13行：

Differential privacy is often discussed in the context of identifying individuals whose information may be in a database. Although it does not directly refer to identification and reidentification attacks, differentially private algorithms probably resist such attacks.

−

粗略地说，如果一个观察者看到一个算法的输出不能判断一个特定个体的信息是否被用于计算，那么这个算法就是有差异的私有的。差分隐私通常是在识别数据库中的个人信息时讨论的。虽然它不直接涉及身份识别和重新身份识别攻击，但差别私有算法可能抵抗这种攻击。<ref name="DMNS06" />

+

简单来说，如果观察者发现某算法的输出不能推断出一个特定个体的信息是否被用于其计算，则该算法是差分隐私的。通常，差分隐私算法会在识别其信息可能存在于数据库中的个体时被讨论。虽然差分隐私算法不直接涉及身份识别和身份重识别攻击，但它或许能够防御这些攻击。<ref name="DMNS06" />

Differential privacy was developed by [[Cryptography|cryptographers]] and thus is often associated with cryptography, and draws much of its language from cryptography.

第19行：第19行：

Differential privacy was developed by cryptographers and thus is often associated with cryptography, and draws much of its language from cryptography.

−

~~差分隐私是由密码学家开发的，因此经常与密码学联系在一起，并从密码学中汲取了大量的语言。~~

+

差分隐私是由密码学家开发的，因此经常与密码学相关，且其大量内容来自于密码学。

==History==

第95行：第95行：

<math>\Pr[\mathcal{A}(D_1) \in S] \leq \exp\left(\epsilon\right) \cdot \Pr[\mathcal{A}(D_2) \in S],</math>

</center>

−

where the probability is taken over the [[randomness]] used by the algorithm.<ref name="DPBook"/>

+

where the probability is taken over the [[randomness]] used by the algorithm.<ref name="DPBook" />

Let ε be a positive real number and \mathcal{A} be a randomized algorithm that takes a dataset as input (representing the actions of the trusted party holding the data).

第112行：第112行：

由于其可组合性、对后处理的鲁棒性以及在相关数据存在时的优雅退化，差分隐私提供了强大而健壮的保证，可以促进模块化设计和差异专用机制的分析。

−

=== Composability ===

+

===Composability===

(Self-)composability refers to the fact that the joint distribution of the outputs of (possibly adaptively chosen) differentially private mechanisms satisfies differential privacy.

第131行：第131行：

平行构图。如果前面的机制是在私有数据库的不相交子集上计算的，那么函数 g 将是(max _ i epsilon _ i)-微分私有。

−

=== Robustness to post-processing ===

+

=== Robustness to post-processing===

For any deterministic or randomized function <math>F</math> defined over the image of the mechanism <math>\mathcal{A}</math>, if <math>\mathcal{A}</math> satisfies ε-differential privacy, so does <math>F(\mathcal{A})</math>.

第144行：第144行：

总之，可组合性和对后期处理的健壮性允许模块化构建和分析不同的私有机制，并激励隐私损失预算的概念。如果访问复杂机制的敏感数据的所有元素都是单独的、不同的私有元素，那么它们的组合也是如此，然后是任意的后处理。

−

=== Group privacy ===

+

===Group privacy===

In general, ε-differential privacy is designed to protect the privacy between neighboring databases which differ only in one row. This means that no adversary with arbitrary auxiliary information can know if '''one''' particular participant submitted his information. However this is also extendable if we want to protect databases differing in <math>c</math> rows, which amounts to adversary with arbitrary auxiliary information can know if '''<math>c</math>''' particular participants submitted their information. This can be achieved because if <math>c</math> items change, the probability dilation is bounded by <math>\exp ( \epsilon c )</math> instead of <math>\exp ( \epsilon )</math>,'''<ref name="Dwork, ICALP 2006" />''' i.e., for D<sub>1</sub> and D<sub>2</sub> differing on <math>c</math> items:

第157行：第157行：

\exp(\epsilon c)\cdot\Pr[\mathcal{A}(D_{2})\in S]\,\!

−

: Pr [ mathcal { a }(d _ {1}) in s ] leq exp (epsilon c) cdot Pr [ mathcal { a }(d _ {2}) in s ] ，！

+

:Pr [ mathcal { a }(d _ {1}) in s ] leq exp (epsilon c) cdot Pr [ mathcal { a }(d _ {2}) in s ] ，！

Thus setting ε instead to <math>\epsilon/c</math> achieves the desired result (protection of <math>c</math> items). In other words, instead of having each item ε-differentially private protected, now every group of <math>c</math> items is ε-differentially private protected (and each item is <math>(\epsilon/c)</math>-differentially private protected).

第165行：第165行：

因此，将 ε 设置为 epsilon/c 可以达到预期的结果(c 项的保护)。换句话说，取代了每个条目 ε- 差别私有保护，现在每组 c 条目都是 ε- 差别私有保护(每个条目 ε/c)-差别私有保护)。

−

== ε-differentially private mechanisms ==

+

==ε-differentially private mechanisms==

Since differential privacy is a probabilistic concept, any differentially private mechanism is necessarily randomized. Some of these, like the Laplace mechanism, described below, rely on adding controlled noise to the function that we want to compute. Others, like the [[Exponential mechanism (differential privacy)|exponential mechanism]]<ref>[http://research.microsoft.com/pubs/65075/mdviadp.pdf F.McSherry and K.Talwar. Mechasim Design via Differential Privacy. Proceedings of the 48th Annual Symposium of Foundations of Computer Science, 2007.]</ref> and posterior sampling<ref>[https://arxiv.org/abs/1306.1066 Christos Dimitrakakis, Blaine Nelson, Aikaterini Mitrokotsa, Benjamin Rubinstein. Robust and Private Bayesian Inference. Algorithmic Learning Theory 2014]</ref> sample from a problem-dependent family of distributions instead.

第172行：第172行：

因为差分隐私是一个概率概念，所以任何差异私有机制都必然是随机的。其中一些，如下面描述的拉普拉斯机制，依赖于在我们要计算的函数中添加受控噪声。其他的，比如指数机制，麦雪莉和塔瓦尔。设计图片来源: 差分隐私。2007年第48届计算机科学基础年会论文集。迪米特拉卡基斯(Dimitrakakis)、布莱恩 · 尼尔森(Blaine Nelson)、艾卡捷里尼 · 米特罗科萨(aikaterina Mitrokotsa)、本杰明 · 鲁宾斯坦(Benjamin Rubinstein)。强大的私人贝叶斯推断。算法学习理论2014年的样本取自一个依赖于问题的分布族。

−

=== Sensitivity ===

+

===Sensitivity===

−

Let <math>d</math> be a positive integer, <math>\mathcal{D}</math> be a collection of datasets, and <math>f \colon \mathcal{D} \rightarrow \mathbb{R}^d</math> be a function. The ''sensitivity'' <ref name="DMNS06"/> of a function, denoted <math>\Delta f</math>, is defined by

+

Let <math>d</math> be a positive integer, <math>\mathcal{D}</math> be a collection of datasets, and <math>f \colon \mathcal{D} \rightarrow \mathbb{R}^d</math> be a function. The ''sensitivity'' <ref name="DMNS06" /> of a function, denoted <math>\Delta f</math>, is defined by

−

: <math>\Delta f=\max \lVert f(D_1)-f(D_2) \rVert_1,</math>

+

:<math>\Delta f=\max \lVert f(D_1)-f(D_2) \rVert_1,</math>

where the maximum is over all pairs of datasets <math>D_1</math> and <math>D_2</math> in <math>\mathcal{D}</math> differing in at most one element and <math>\lVert \cdot \rVert_1</math> denotes the [[Taxicab geometry|<math>\ell_1</math> norm]].

Let d be a positive integer, \mathcal{D} be a collection of datasets, and f \colon \mathcal{D} \rightarrow \mathbb{R}^d be a function. The sensitivity of a function, denoted \Delta f, is defined by

−

: \Delta f=\max \lVert f(D_1)-f(D_2) \rVert_1,

+

:\Delta f=\max \lVert f(D_1)-f(D_2) \rVert_1,

where the maximum is over all pairs of datasets D_1 and D_2 in \mathcal{D} differing in at most one element and \lVert \cdot \rVert_1 denotes the \ell_1 norm.

第195行：第195行：

有一些技术(下面将描述) ，我们可以使用这些技术为低灵敏度函数创建一个差分私有算法。

−

=== The Laplace mechanism ===

+

===The Laplace mechanism===

The Laplace mechanism adds Laplace noise (i.e. noise from the [[Laplace distribution]], which can be expressed by probability density function <math>\text{noise}(y)\propto \exp(-|y|/\lambda)\,\!</math>, which has mean zero and standard deviation <math>\sqrt{2} \lambda\,\!</math>). Now in our case we define the output function of <math>\mathcal{A}\,\!</math> as a real valued function (called as the transcript output by <math>\mathcal{A}\,\!</math>) as <math>\mathcal{T}_{\mathcal{A}}(x)=f(x)+Y\,\!</math> where <math>Y \sim \text{Lap}(\lambda)\,\!\,\!</math> and <math>f\,\!</math> is the original real valued query/function we planned to execute on the database. Now clearly <math>\mathcal{T}_{\mathcal{A}}(x)\,\!</math> can be considered to be a continuous random variable, where

第201行：第201行： −

The Laplace mechanism adds Laplace noise (i.e. noise from the Laplace distribution, which can be expressed by probability density function \text{noise}(y)\propto \exp(-|y|/\lambda)\,\!, which has mean zero and standard deviation \sqrt{2} \lambda\,\!). Now in our case we define the output function of \mathcal{A}\,\! as a real valued function (called as the transcript output by \mathcal{A}\,\!) as \mathcal{T}_{\mathcal{A}}(x)=f(x)+Y\,\! where Y \sim \text{Lap}(\lambda)\,\!\,\! and f\,\! is the original real valued query/function we planned to execute on the database. Now clearly \mathcal{T}_{\mathcal{A}}(x)\,\! can be considered to be a continuous random variable, where

+

<nowiki>The Laplace mechanism adds Laplace noise (i.e. noise from the Laplace distribution, which can be expressed by probability density function \text{noise}(y)\propto \exp(-|y|/\lambda)\,\!, which has mean zero and standard deviation \sqrt{2} \lambda\,\!). Now in our case we define the output function of \mathcal{A}\,\! as a real valued function (called as the transcript output by \mathcal{A}\,\!) as \mathcal{T}_{\mathcal{A}}(x)=f(x)+Y\,\! where Y \sim \text{Lap}(\lambda)\,\!\,\! and f\,\! is the original real valued query/function we planned to execute on the database. Now clearly \mathcal{T}_{\mathcal{A}}(x)\,\! can be considered to be a continuous random variable, where</nowiki>

−

拉普拉斯机制增加了拉普拉斯的噪音(即。拉普拉斯分布的噪声，可以用概率密度函数文本{ noise }(y) propto exp (- | y |/lambda)来表示，！这个函数的平均值是0和标准差的 sqrt {2} lambda，!).现在，在我们的例子中，我们定义了数学{ a }的输出函数，！作为一个实值函数(由 mathcal { a } ，!)数学{ t } _ {数学{ a }}(x) = f (x) + y，！其中 y sim 文本{ Lap }(lambda) ，! ，！还有 f！是我们计划在数据库上执行的原始实值查询/函数。现在清楚的数学{ t } _ {数学{ a }(x) ，！可以被认为是一个连续的随机变量

+

<nowiki>拉普拉斯机制增加了拉普拉斯的噪音(即。拉普拉斯分布的噪声，可以用概率密度函数文本{ noise }(y) propto exp (- | y |/lambda)来表示，！这个函数的平均值是0和标准差的 sqrt {2} lambda，!).现在，在我们的例子中，我们定义了数学{ a }的输出函数，！作为一个实值函数(由 mathcal { a } ，!)数学{ t } _ {数学{ a }}(x) = f (x) + y，！其中 y sim 文本{ Lap }(lambda) ，! ，！还有 f！是我们计划在数据库上执行的原始实值查询/函数。现在清楚的数学{ t } _ {数学{ a }(x) ，！可以被认为是一个连续的随机变量</nowiki>

:<math>\frac{\mathrm{pdf}(\mathcal{T}_{\mathcal{A},D_1}(x)=t)}{\mathrm{pdf}(\mathcal{T}_{\mathcal{A},D_2}(x)=t)}=\frac{\text{noise}(t-f(D_1))}{\text{noise}(t-f(D_2))}\,\!</math>

第212行：第212行：

−

: frc { mathrm { pdf }(mathcal { t } _ { mathcal { a } ，d _ 1}(x) = t)}{ mathrm { pdf }(mathcal { t } _ { mathcal { a } ，d _ 2}(x) = t)} = frc { text { noise }(t-f (d _ 1))}{ text { noise }(t-f (d _ 2))} ，！

+

:frc { mathrm { pdf }(mathcal { t } _ { mathcal { a } ，d _ 1}(x) = t)}{ mathrm { pdf }(mathcal { t } _ { mathcal { a } ，d _ 2}(x) = t)} = frc { text { noise }(t-f (d _ 1))}{ text { noise }(t-f (d _ 2))} ，！

which is at most <math>e^{\frac{|f(D_{1})-f(D_{2})|}{\lambda}}\leq e^{\frac{\Delta(f)}{\lambda}}\,\!</math>. We can consider <math>\frac{\Delta(f)}{\lambda}\,\!</math> to be the privacy factor <math>\epsilon\,\!</math>. Thus <math>\mathcal{T}\,\!</math> follows a differentially private mechanism (as can be seen from [[#ε-differential privacy|the definition above]]). If we try to use this concept in our diabetes example then it follows from the above derived fact that in order to have <math>\mathcal{A}\,\!</math> as the <math>\epsilon\,\!</math>-differential private algorithm we need to have <math>\lambda=1/\epsilon\,\!</math>. Though we have used Laplace noise here, other forms of noise, such as the Gaussian Noise, can be employed, but they may require a slight relaxation of the definition of differential privacy.<ref name="Dwork, ICALP 2006" />

−

which is at most e^{\frac{|f(D_{1})-f(D_{2})|}{\lambda}}\leq e^{\frac{\Delta(f)}{\lambda}}\,\!. We can consider \frac{\Delta(f)}{\lambda}\,\! to be the privacy factor \epsilon\,\!. Thus \mathcal{T}\,\! follows a differentially private mechanism (as can be seen from the definition above). If we try to use this concept in our diabetes example then it follows from the above derived fact that in order to have \mathcal{A}\,\! as the \epsilon\,\!-differential private algorithm we need to have \lambda=1/\epsilon\,\!. Though we have used Laplace noise here, other forms of noise, such as the Gaussian Noise, can be employed, but they may require a slight relaxation of the definition of differential privacy.

+

<nowiki>which is at most e^{\frac{|f(D_{1})-f(D_{2})|}{\lambda}}\leq e^{\frac{\Delta(f)}{\lambda}}\,\!. We can consider \frac{\Delta(f)}{\lambda}\,\! to be the privacy factor \epsilon\,\!. Thus \mathcal{T}\,\! follows a differentially private mechanism (as can be seen from the definition above). If we try to use this concept in our diabetes example then it follows from the above derived fact that in order to have \mathcal{A}\,\! as the \epsilon\,\!-differential private algorithm we need to have \lambda=1/\epsilon\,\!. Though we have used Laplace noise here, other forms of noise, such as the Gaussian Noise, can be employed, but they may require a slight relaxation of the definition of differential privacy.</nowiki>

最多 e ^ { frac { | f (d _ {1})-f (d _ {2}) | }{ lambda } leq e ^ { frac { Delta (f)}{ lambda } ，。我们可以考虑 frac { Delta (f)}{ lambda } ，！成为隐私保护因素。因此数学{ t } ，！遵循不同的私有机制(从上面的定义可以看出)。如果我们在我们的糖尿病示例中尝试使用这个概念，那么从上面的派生事实可以推出，为了得到数学{ a } ，！作为 epsilon!-我们需要 lambda = 1/epsilon! 。虽然我们在这里使用了拉普拉斯噪音，但是也可以使用其他形式的噪音，比如高斯噪音，但是它们可能需要稍微放宽差分隐私的定义。

第234行：第234行：

{| class="wikitable" style="margin-left: auto; margin-right: auto; border: none;"

|-

−

! Name !! Has Diabetes (X)

+

!Name!!Has Diabetes (X)

|-

−

| Ross

+

|Ross

−

|| 1

+

||1

|-

| Monica

−

|| 1

+

||1

|-

−

| Joey

+

|Joey

−

|| 0

+

||0

|-

−

| Phoebe

+

|Phoebe

−

|| 0

+

||0

|-

−

| Chandler

+

|Chandler

−

|| 1

+

||1

|-

−

| Rachel

+

|Rachel

−

|| 0

+

||0

|}

{| class="wikitable" style="margin-left: auto; margin-right: auto; border: none;"

|-

−

! Name !! Has Diabetes (X)

+

!Name!!Has Diabetes (X)

|-

−

| Ross

+

|Ross

−

|| 1

+

||1

|-

| Monica

−

|| 1

+

||1

|-

−

| Joey

+

|Joey

−

|| 0

+

||0

|-

−

| Phoebe

+

|Phoebe

−

|| 0

+

||0

|-

−

| Chandler

+

|Chandler

−

|| 1

+

||1

|-

−

| Rachel

+

|Rachel

−

|| 0

+

||0

|}

{| class="wikitable" style="margin-left: auto; margin-right: auto; border: none;"

|-

−

!姓名! ！患有糖尿病(x) |<nowiki>-| Ross | | 1 |-| Monica | | | 1 | |-| Joey | | 0 |-| Phoebe | 0 |-|-| Chandler | | 1 |-|-| Rachel | | | | |</nowiki>

+

! 姓名! ！患有糖尿病(x) |<nowiki>-| Ross | | 1 |-| Monica | | | 1 | |-| Joey | | 0 |-| Phoebe | 0 |-|-| Chandler | | 1 |-|-| Rachel | | | | |</nowiki>

Now suppose a malicious user (often termed an ''adversary'') wants to find whether Chandler has diabetes or not. Suppose he also knows in which row of the database Chandler resides. Now suppose the adversary is only allowed to use a particular form of query <math>Q_i</math> that returns the partial sum of the first <math>i</math> rows of column <math>X</math> in the database. In order to find Chandler's diabetes status the adversary executes <math>Q_5(D_1)</math> and <math>Q_4(D_1)</math>, then computes their difference. In this example, <math>Q_5(D_1) = 3</math> and <math>Q_4(D_1) = 2</math>, so their difference is 1. This indicates that the "Has Diabetes" field in Chandler's row must be 1. This example highlights how individual information can be compromised even without explicitly querying for the information of a specific individual.

第294行：第294行：

继续这个例子，如果我们用(Chandler，1)替换(Chandler，0)来构造 d2，那么这个恶意的对手将能够通过计算每个数据集的 q _ 5-q _ 4来区分 d _ 2和 d _ 1。如果对手被要求通过一个 epsilon-differentially private 算法接收值 q _ i，对于一个足够小的 epsilon，那么他或她将无法区分两个数据集。

−

===Randomized response ===

+

===Randomized response===

第303行：第303行：

一个简单的例子，尤其是在社会科学领域，就是让一个人回答“你拥有属性 a 吗?”？”，根据下列程序:

−

# [[Coin flipping|Toss a coin]].

+

#[[Coin flipping|Toss a coin]].

# If heads, then toss the coin again (ignoring the outcome), and answer the question honestly.

−

# If tails, then toss the coin again and answer "Yes" if heads, "No" if tails.

+

#If tails, then toss the coin again and answer "Yes" if heads, "No" if tails.

−

# Toss a coin.

+

#Toss a coin.

# If heads, then toss the coin again (ignoring the outcome), and answer the question honestly.

−

# If tails, then toss the coin again and answer "Yes" if heads, "No" if tails.

+

#If tails, then toss the coin again and answer "Yes" if heads, "No" if tails.

# 抛硬币。# 如果是正面，再掷硬币(忽略结果) ，诚实地回答问题。# 如果是反面，再掷一次硬币，如果是正面，回答“是”; 如果是反面，回答“否”。

第339行：第339行：

虽然这个例子受到了随机化回答的启发，可能适用于微数据(例如，发布每个响应的数据集) ，但根据定义，差分隐私排除了微数据发布，并且只适用于查询(例如，将单个响应聚合成一个结果) ，因为这将违反要求，更具体地说，是一个主题参与或不参与的似是而非的否认。辛西娅。“为私人数据分析奠定坚实的基础。”美国计算机学会通讯54.1(2011) : 86-95，上注19，第91页. Bambauer，Jane，Krishnamurty Muralidhar，and Rathindra Sarathy。“愚人的黄金: 对差分隐私的插图式批评。”Vand.J. Ent.北京科技发展有限公司。L. 16(2013) : 701.

−

=== Stable transformations ===

+

=== Stable transformations===

−

A transformation <math>T</math> is <math>c</math>-stable if the [[Hamming distance]] between <math>T(A)</math> and <math>T(B)</math> is at most <math>c</math>-times the Hamming distance between <math>A</math> and <math>B</math> for any two databases <math>A,B</math>. Theorem 2 in <ref name="PINQ"/> asserts that if there is a mechanism <math>M</math> that is <math>\epsilon</math>-differentially private, then the composite mechanism <math>M\circ T</math> is <math>(\epsilon \times c)</math>-differentially private.

+

A transformation <math>T</math> is <math>c</math>-stable if the [[Hamming distance]] between <math>T(A)</math> and <math>T(B)</math> is at most <math>c</math>-times the Hamming distance between <math>A</math> and <math>B</math> for any two databases <math>A,B</math>. Theorem 2 in <ref name="PINQ" /> asserts that if there is a mechanism <math>M</math> that is <math>\epsilon</math>-differentially private, then the composite mechanism <math>M\circ T</math> is <math>(\epsilon \times c)</math>-differentially private.

A transformation T is c-stable if the Hamming distance between T(A) and T(B) is at most c-times the Hamming distance between A and B for any two databases A,B. Theorem 2 in asserts that if there is a mechanism M that is \epsilon-differentially private, then the composite mechanism M\circ T is (\epsilon \times c)-differentially private.

第355行：第355行：

==Other notions of differential privacy==

−

Since differential privacy is considered to be too strong or weak for some applications, many versions of it have been proposed.<ref name="DP19"/> The most widespread relaxation is (ε, δ)-differential privacy,<ref name="DKMMN06"/> which weakens the definition by allowing an additional small δ density of probability on which the upper bound ε does not hold.

+

Since differential privacy is considered to be too strong or weak for some applications, many versions of it have been proposed.<ref name="DP19" /> The most widespread relaxation is (ε, δ)-differential privacy,<ref name="DKMMN06" /> which weakens the definition by allowing an additional small δ density of probability on which the upper bound ε does not hold.

Since differential privacy is considered to be too strong or weak for some applications, many versions of it have been proposed. The most widespread relaxation is (ε, δ)-differential privacy, which weakens the definition by allowing an additional small δ density of probability on which the upper bound ε does not hold.

第361行：第361行：

由于对于某些应用程序来说，差分隐私被认为太强或太弱，因此人们提出了许多版本。最广泛的松弛是(ε，δ)-差分隐私，它通过允许增加一个上限 ε 不成立的概率密度 δ 来削弱定义。

−

== Adoption of differential privacy in real-world applications ==

+

==Adoption of differential privacy in real-world applications==

Several uses of differential privacy in practice are known to date:

−

* 2008: [[United States Census Bureau|U.S. Census Bureau]], for showing commuting patterns.<ref name="MachanavajjhalaKAGV08"/>

+

*2008: [[United States Census Bureau|U.S. Census Bureau]], for showing commuting patterns.<ref name="MachanavajjhalaKAGV08" />

−

* 2014: [[Google]]'s RAPPOR, for telemetry such as learning statistics about unwanted software hijacking users' settings. <ref name="RAPPOR"/><ref>{{Citation|title=google/rappor|date=2021-07-15|url=https://github.com/google/rappor|publisher=GitHub}}</ref>

+

*2014: [[Google]]'s RAPPOR, for telemetry such as learning statistics about unwanted software hijacking users' settings. <ref name="RAPPOR" /><ref>{{Citation|title=google/rappor|date=2021-07-15|url=https://github.com/google/rappor|publisher=GitHub}}</ref>

−

* 2015: Google, for sharing historical traffic statistics.<ref name="Eland"/>

+

*2015: Google, for sharing historical traffic statistics.<ref name="Eland" />

−

* 2016: [[Apple Inc.|Apple]] announced its intention to use differential privacy in [[iOS 10]] to improve its [[Intelligent personal assistant]] technology.<ref>{{cite web|title=Apple - Press Info - Apple Previews iOS 10, the Biggest iOS Release Ever|url=https://www.apple.com/pr/library/2016/06/13Apple-Previews-iOS-10-The-Biggest-iOS-Release-Ever.html|website=Apple|access-date=16 June 2016}}</ref>

+

*2016: [[Apple Inc.|Apple]] announced its intention to use differential privacy in [[iOS 10]] to improve its [[Intelligent personal assistant]] technology.<ref>{{cite web|title=Apple - Press Info - Apple Previews iOS 10, the Biggest iOS Release Ever|url=https://www.apple.com/pr/library/2016/06/13Apple-Previews-iOS-10-The-Biggest-iOS-Release-Ever.html|website=Apple|access-date=16 June 2016}}</ref>

−

* 2017: Microsoft, for telemetry in Windows.<ref name="DpWinTelemetry"/>

+

*2017: Microsoft, for telemetry in Windows.<ref name="DpWinTelemetry" />

−

* 2019: Privitar Lens is an API using differential privacy.<ref>{{cite web|title=Privitar Lens|url=https://www.privitar.com/privitar-lens|access-date=20 February 2018}}</ref>

+

*2019: Privitar Lens is an API using differential privacy.<ref>{{cite web|title=Privitar Lens|url=https://www.privitar.com/privitar-lens|access-date=20 February 2018}}</ref>

−

* 2020: LinkedIn, for advertiser queries.<ref name="DpLinkedIn"/>

+

*2020: LinkedIn, for advertiser queries.<ref name="DpLinkedIn" />

Several uses of differential privacy in practice are known to date:

−

* 2008: U.S. Census Bureau, for showing commuting patterns.

+

*2008: U.S. Census Bureau, for showing commuting patterns.

−

* 2014: Google's RAPPOR, for telemetry such as learning statistics about unwanted software hijacking users' settings.

+

*2014: Google's RAPPOR, for telemetry such as learning statistics about unwanted software hijacking users' settings.

* 2015: Google, for sharing historical traffic statistics.

−

* 2016: Apple announced its intention to use differential privacy in iOS 10 to improve its Intelligent personal assistant technology.

+

*2016: Apple announced its intention to use differential privacy in iOS 10 to improve its Intelligent personal assistant technology.

−

* 2017: Microsoft, for telemetry in Windows.

+

*2017: Microsoft, for telemetry in Windows.

−

* 2019: Privitar Lens is an API using differential privacy.

+

*2019: Privitar Lens is an API using differential privacy.

* 2020: LinkedIn, for advertiser queries.

2008: u.s. Census Bureau，for shows comforting patterns. 在实践中，差分隐私的几个用途已经为人所知:

−

* 2008: 美国人口普查局，显示通勤模式。

+

*2008: 美国人口普查局，显示通勤模式。

−

* 2014年: 谷歌的 RAPPOR，用于遥测，例如了解不受欢迎的软件劫持用户设置的统计数据。2015: Google，分享历史流量统计数据。

+

*2014年: 谷歌的 RAPPOR，用于遥测，例如了解不受欢迎的软件劫持用户设置的统计数据。2015: Google，分享历史流量统计数据。

−

* 2016年: 苹果公司宣布打算在 iOS 10中使用差分隐私智能个人助理来改进其智能个人助理技术。

+

*2016年: 苹果公司宣布打算在 iOS 10中使用差分隐私智能个人助理来改进其智能个人助理技术。

−

* 2017: 微软，Windows 遥测系统。2019: priveritar Lens 是一个使用差分隐私的 API。2020: LinkedIn，for advertiser queries.

+

*2017: 微软，Windows 遥测系统。2019: priveritar Lens 是一个使用差分隐私的 API。2020: LinkedIn，for advertiser queries.

−

== Public purpose considerations ==

+

==Public purpose considerations==

There are several public purpose considerations regarding differential privacy that are important to consider, especially for policymakers and policy-focused audiences interested in the social opportunities and risks of the technology:<ref>{{Cite web|title=Technology Factsheet: Differential Privacy|url=https://www.belfercenter.org/publication/technology-factsheet-differential-privacy|access-date=2021-04-12|website=Belfer Center for Science and International Affairs|language=en}}</ref>

第395行：第395行：

关于差分隐私技术，有几个公共目的方面的考虑是需要考虑的，特别是对于那些对技术的社会机遇和风险感兴趣的决策者和政策关注的受众:

−

* '''Data Utility & Accuracy.''' The main concern with differential privacy is the tradeoff between data utility and individual privacy. If the privacy loss parameter is set to favor utility, the privacy benefits are lowered (less “noise” is injected into the system); if the privacy loss parameter is set to favor heavy privacy, the accuracy and utility of the dataset are lowered (more “noise” is injected into the system). It is important for policymakers to consider the tradeoffs posed by differential privacy in order to help set appropriate best practices and standards around the use of this privacy preserving practice, especially considering the diversity in organizational use cases. It is worth noting, though, that decreased accuracy and utility is a common issue among all statistical disclosure limitation methods and is not unique to differential privacy. What is unique, however, is how policymakers, researchers, and implementers can consider mitigating against the risks presented through this tradeoff.

+

*'''Data Utility & Accuracy.''' The main concern with differential privacy is the tradeoff between data utility and individual privacy. If the privacy loss parameter is set to favor utility, the privacy benefits are lowered (less “noise” is injected into the system); if the privacy loss parameter is set to favor heavy privacy, the accuracy and utility of the dataset are lowered (more “noise” is injected into the system). It is important for policymakers to consider the tradeoffs posed by differential privacy in order to help set appropriate best practices and standards around the use of this privacy preserving practice, especially considering the diversity in organizational use cases. It is worth noting, though, that decreased accuracy and utility is a common issue among all statistical disclosure limitation methods and is not unique to differential privacy. What is unique, however, is how policymakers, researchers, and implementers can consider mitigating against the risks presented through this tradeoff.

−

* Data Utility & Accuracy. The main concern with differential privacy is the tradeoff between data utility and individual privacy. If the privacy loss parameter is set to favor utility, the privacy benefits are lowered (less “noise” is injected into the system); if the privacy loss parameter is set to favor heavy privacy, the accuracy and utility of the dataset are lowered (more “noise” is injected into the system). It is important for policymakers to consider the tradeoffs posed by differential privacy in order to help set appropriate best practices and standards around the use of this privacy preserving practice, especially considering the diversity in organizational use cases. It is worth noting, though, that decreased accuracy and utility is a common issue among all statistical disclosure limitation methods and is not unique to differential privacy. What is unique, however, is how policymakers, researchers, and implementers can consider mitigating against the risks presented through this tradeoff.

+

*Data Utility & Accuracy. The main concern with differential privacy is the tradeoff between data utility and individual privacy. If the privacy loss parameter is set to favor utility, the privacy benefits are lowered (less “noise” is injected into the system); if the privacy loss parameter is set to favor heavy privacy, the accuracy and utility of the dataset are lowered (more “noise” is injected into the system). It is important for policymakers to consider the tradeoffs posed by differential privacy in order to help set appropriate best practices and standards around the use of this privacy preserving practice, especially considering the diversity in organizational use cases. It is worth noting, though, that decreased accuracy and utility is a common issue among all statistical disclosure limitation methods and is not unique to differential privacy. What is unique, however, is how policymakers, researchers, and implementers can consider mitigating against the risks presented through this tradeoff.

−

* 数据的实用性及准确性。差分隐私的主要关注点在于数据效用和个人隐私之间的权衡。如果将隐私损失参数设置为有利于实用性，则隐私好处降低(向系统中注入的“噪音”较少) ; 如果将隐私损失参数设置为有利于重隐私性，则数据集的准确性和实用性降低(向系统中注入更多的“噪音”)。对于决策者来说，重要的是要考虑到差分隐私的权衡，以帮助建立适当的最佳实践和标准来使用这种隐私保护实践，特别是考虑到组织用例的多样性。值得注意的是，在所有的统计披露限制方法中，降低准确性和效用是一个共同的问题，并不是差分隐私唯一的。然而，独特之处在于，决策者、研究人员和实施者可以考虑如何减轻这种权衡带来的风险。

+

*数据的实用性及准确性。差分隐私的主要关注点在于数据效用和个人隐私之间的权衡。如果将隐私损失参数设置为有利于实用性，则隐私好处降低(向系统中注入的“噪音”较少) ; 如果将隐私损失参数设置为有利于重隐私性，则数据集的准确性和实用性降低(向系统中注入更多的“噪音”)。对于决策者来说，重要的是要考虑到差分隐私的权衡，以帮助建立适当的最佳实践和标准来使用这种隐私保护实践，特别是考虑到组织用例的多样性。值得注意的是，在所有的统计披露限制方法中，降低准确性和效用是一个共同的问题，并不是差分隐私唯一的。然而，独特之处在于，决策者、研究人员和实施者可以考虑如何减轻这种权衡带来的风险。

−

* '''Data Privacy & Security.''' Differential privacy provides a quantified measure of privacy loss and an upper bound and allows curators to choose the explicit tradeoff between privacy and accuracy. It is robust to still unknown privacy attacks. However, it encourages greater data sharing, which if done poorly, increases privacy risk. Differential privacy implies that privacy is protected, but this depends very much on the privacy loss parameter chosen and may instead lead to a false sense of security. Finally, though it is robust against unforeseen future privacy attacks, a countermeasure may be devised that we cannot predict.

+

*'''Data Privacy & Security.''' Differential privacy provides a quantified measure of privacy loss and an upper bound and allows curators to choose the explicit tradeoff between privacy and accuracy. It is robust to still unknown privacy attacks. However, it encourages greater data sharing, which if done poorly, increases privacy risk. Differential privacy implies that privacy is protected, but this depends very much on the privacy loss parameter chosen and may instead lead to a false sense of security. Finally, though it is robust against unforeseen future privacy attacks, a countermeasure may be devised that we cannot predict.

−

* Data Privacy & Security. Differential privacy provides a quantified measure of privacy loss and an upper bound and allows curators to choose the explicit tradeoff between privacy and accuracy. It is robust to still unknown privacy attacks. However, it encourages greater data sharing, which if done poorly, increases privacy risk. Differential privacy implies that privacy is protected, but this depends very much on the privacy loss parameter chosen and may instead lead to a false sense of security. Finally, though it is robust against unforeseen future privacy attacks, a countermeasure may be devised that we cannot predict.

+

*Data Privacy & Security. Differential privacy provides a quantified measure of privacy loss and an upper bound and allows curators to choose the explicit tradeoff between privacy and accuracy. It is robust to still unknown privacy attacks. However, it encourages greater data sharing, which if done poorly, increases privacy risk. Differential privacy implies that privacy is protected, but this depends very much on the privacy loss parameter chosen and may instead lead to a false sense of security. Finally, though it is robust against unforeseen future privacy attacks, a countermeasure may be devised that we cannot predict.

−

* 资料私隐及保安。差分隐私图书馆提供了一个量化的隐私损失度量和上限，并允许馆长在隐私和准确性之间做出明确的权衡。它对仍然未知的隐私攻击是健壮的。然而，它鼓励更大的数据共享，如果做得不好，会增加隐私风险。差分隐私意味着隐私是受到保护的，但这在很大程度上取决于选择的隐私损失参数，并可能会导致错误的安全感。最后，尽管它对未来不可预见的隐私攻击是健壮的，但可以设计出一种我们无法预测的对策。

+

*资料私隐及保安。差分隐私图书馆提供了一个量化的隐私损失度量和上限，并允许馆长在隐私和准确性之间做出明确的权衡。它对仍然未知的隐私攻击是健壮的。然而，它鼓励更大的数据共享，如果做得不好，会增加隐私风险。差分隐私意味着隐私是受到保护的，但这在很大程度上取决于选择的隐私损失参数，并可能会导致错误的安全感。最后，尽管它对未来不可预见的隐私攻击是健壮的，但可以设计出一种我们无法预测的对策。

==See also==

第419行：第419行：

*Exponential mechanism (differential privacy) – a technique for designing differentially private algorithms

*k-anonymity

−

*Differentially private analysis of graphs

+

* Differentially private analysis of graphs

*Protected health information

−

* 准标识符

+

*准标识符

−

* 指数机制(差分隐私)-一种设计不同私有算法的技术

+

*指数机制(差分隐私)-一种设计不同私有算法的技术

−

* k-匿名

+

*k-匿名

−

* 图的不同私有分析

+

*图的不同私有分析

−

* 受保护的健康信息

+

*受保护的健康信息

==References==

第531行：第531行：

==Further reading==

−

*[https://desfontain.es/privacy/index.html A reading list on differential privacy]

+

*[https://desfontain.es/privacy/index.html A reading list on differential privacy]

*[https://journalprivacyconfidentiality.org/index.php/jpc/article/view/404 Abowd, John. 2017. “How Will Statistical Agencies Operate When All Data Are Private?”. Journal of Privacy and Confidentiality 7 (3).] {{doi|10.29012/jpc.v7i3.404}} ([https://www2.census.gov/cac/sac/meetings/2017-09/role-statistical-agency.pdf slides])

−

* [http://www.jetlaw.org/wp-content/uploads/2018/12/4_Wood_Final.pdf "Differential Privacy: A Primer for a Non-technical Audience"], Kobbi Nissim, Thomas Steinke, Alexandra Wood, [[Micah Altman]], Aaron Bembenek, Mark Bun, Marco Gaboardi, David R. O’Brien, and Salil Vadhan, Harvard Privacy Tools Project, February 14, 2018

+

*[http://www.jetlaw.org/wp-content/uploads/2018/12/4_Wood_Final.pdf "Differential Privacy: A Primer for a Non-technical Audience"], Kobbi Nissim, Thomas Steinke, Alexandra Wood, [[Micah Altman]], Aaron Bembenek, Mark Bun, Marco Gaboardi, David R. O’Brien, and Salil Vadhan, Harvard Privacy Tools Project, February 14, 2018

−

* Dinur, Irit and Kobbi Nissim. 2003. Revealing information while preserving privacy. In Proceedings of the twenty-second ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems(PODS '03). ACM, New York, NY, USA, 202-210. {{doi|10.1145/773153.773173}}.

+

*Dinur, Irit and Kobbi Nissim. 2003. Revealing information while preserving privacy. In Proceedings of the twenty-second ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems(PODS '03). ACM, New York, NY, USA, 202-210. {{doi|10.1145/773153.773173}}.

−

* Dwork, Cynthia, Frank McSherry, Kobbi Nissim, and Adam Smith. 2006. in Halevi, S. & Rabin, T. (Eds.) Calibrating Noise to Sensitivity in Private Data Analysis Theory of Cryptography: Third Theory of Cryptography Conference, TCC 2006, New York, NY, USA, March 4–7, 2006. Proceedings, Springer Berlin Heidelberg, 265-284, {{doi|10.1007/11681878 14}}.

+

*Dwork, Cynthia, Frank McSherry, Kobbi Nissim, and Adam Smith. 2006. in Halevi, S. & Rabin, T. (Eds.) Calibrating Noise to Sensitivity in Private Data Analysis Theory of Cryptography: Third Theory of Cryptography Conference, TCC 2006, New York, NY, USA, March 4–7, 2006. Proceedings, Springer Berlin Heidelberg, 265-284, {{doi|10.1007/11681878 14}}.

−

* Dwork, Cynthia. 2006. Differential Privacy, 33rd International Colloquium on Automata, Languages and Programming, part II (ICALP 2006), Springer Verlag, 4052, 1-12, {{ISBN|3-540-35907-9}}.

+

*Dwork, Cynthia. 2006. Differential Privacy, 33rd International Colloquium on Automata, Languages and Programming, part II (ICALP 2006), Springer Verlag, 4052, 1-12, {{ISBN|3-540-35907-9}}.

−

* Dwork, Cynthia and Aaron Roth. 2014. The Algorithmic Foundations of Differential Privacy. Foundations and Trends in Theoretical Computer Science. Vol. 9, Nos. 3–4. 211–407, {{doi|10.1561/0400000042}}.

+

*Dwork, Cynthia and Aaron Roth. 2014. The Algorithmic Foundations of Differential Privacy. Foundations and Trends in Theoretical Computer Science. Vol. 9, Nos. 3–4. 211–407, {{doi|10.1561/0400000042}}.

−

* Machanavajjhala, Ashwin, Daniel Kifer, John M. Abowd, Johannes Gehrke, and Lars Vilhuber. 2008. Privacy: Theory Meets Practice on the Map, International Conference on Data Engineering (ICDE) 2008: 277-286, {{doi|10.1109/ICDE.2008.4497436}}.

+

*Machanavajjhala, Ashwin, Daniel Kifer, John M. Abowd, Johannes Gehrke, and Lars Vilhuber. 2008. Privacy: Theory Meets Practice on the Map, International Conference on Data Engineering (ICDE) 2008: 277-286, {{doi|10.1109/ICDE.2008.4497436}}.

* Dwork, Cynthia and Moni Naor. 2010. On the Difficulties of Disclosure Prevention in Statistical Databases or The Case for Differential Privacy, Journal of Privacy and Confidentiality: Vol. 2: Iss. 1, Article 8. Available at: http://repository.cmu.edu/jpc/vol2/iss1/8.

−

* Kifer, Daniel and Ashwin Machanavajjhala. 2011. No free lunch in data privacy. In Proceedings of the 2011 ACM SIGMOD International Conference on Management of data (SIGMOD '11). ACM, New York, NY, USA, 193-204. {{doi|10.1145/1989323.1989345}}.

+

*Kifer, Daniel and Ashwin Machanavajjhala. 2011. No free lunch in data privacy. In Proceedings of the 2011 ACM SIGMOD International Conference on Management of data (SIGMOD '11). ACM, New York, NY, USA, 193-204. {{doi|10.1145/1989323.1989345}}.

−

* Erlingsson, Úlfar, Vasyl Pihur and Aleksandra Korolova. 2014. RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response. In Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security (CCS '14). ACM, New York, NY, USA, 1054-1067. {{doi|10.1145/2660267.2660348}}.

+

*Erlingsson, Úlfar, Vasyl Pihur and Aleksandra Korolova. 2014. RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response. In Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security (CCS '14). ACM, New York, NY, USA, 1054-1067. {{doi|10.1145/2660267.2660348}}.

−

* Abowd, John M. and Ian M. Schmutte. 2017 . Revisiting the economics of privacy: Population statistics and confidentiality protection as public goods. Labor Dynamics Institute, Cornell University, Labor Dynamics Institute, Cornell University, at https://digitalcommons.ilr.cornell.edu/ldi/37/

+

*Abowd, John M. and Ian M. Schmutte. 2017 . Revisiting the economics of privacy: Population statistics and confidentiality protection as public goods. Labor Dynamics Institute, Cornell University, Labor Dynamics Institute, Cornell University, at https://digitalcommons.ilr.cornell.edu/ldi/37/

−

* Abowd, John M. and Ian M. Schmutte. Forthcoming. An Economic Analysis of Privacy Protection and Statistical Accuracy as Social Choices. American Economic Review, {{arxiv|1808.06303}}

+

*Abowd, John M. and Ian M. Schmutte. Forthcoming. An Economic Analysis of Privacy Protection and Statistical Accuracy as Social Choices. American Economic Review, {{arxiv|1808.06303}}

−

* Apple, Inc. 2016. Apple previews iOS 10, the biggest iOS release ever. Press Release (June 13). https://www.apple.com/newsroom/2016/06/apple-previews-ios-10-biggest-ios-release-ever.html.

+

*Apple, Inc. 2016. Apple previews iOS 10, the biggest iOS release ever. Press Release (June 13). https://www.apple.com/newsroom/2016/06/apple-previews-ios-10-biggest-ios-release-ever.html.

* Ding, Bolin, Janardhan Kulkarni, and Sergey Yekhanin 2017. Collecting Telemetry Data Privately, NIPS 2017.

−

* http://www.win-vector.com/blog/2015/10/a-simpler-explanation-of-differential-privacy/

+

*http://www.win-vector.com/blog/2015/10/a-simpler-explanation-of-differential-privacy/

−

* Ryffel, Theo, Andrew Trask, et. al. [[arxiv:1811.04017|"A generic framework for privacy preserving deep learning"]]

+

*Ryffel, Theo, Andrew Trask, et. al. [[arxiv:1811.04017|"A generic framework for privacy preserving deep learning"]]

−

*A reading list on differential privacy

+

* A reading list on differential privacy

−

*Abowd, John. 2017. “How Will Statistical Agencies Operate When All Data Are Private?”. Journal of Privacy and Confidentiality 7 (3). (slides)

+

* Abowd, John. 2017. “How Will Statistical Agencies Operate When All Data Are Private?”. Journal of Privacy and Confidentiality 7 (3). (slides)

−

* "Differential Privacy: A Primer for a Non-technical Audience", Kobbi Nissim, Thomas Steinke, Alexandra Wood, Micah Altman, Aaron Bembenek, Mark Bun, Marco Gaboardi, David R. O’Brien, and Salil Vadhan, Harvard Privacy Tools Project, February 14, 2018

+

*"Differential Privacy: A Primer for a Non-technical Audience", Kobbi Nissim, Thomas Steinke, Alexandra Wood, Micah Altman, Aaron Bembenek, Mark Bun, Marco Gaboardi, David R. O’Brien, and Salil Vadhan, Harvard Privacy Tools Project, February 14, 2018

−

* Dinur, Irit and Kobbi Nissim. 2003. Revealing information while preserving privacy. In Proceedings of the twenty-second ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems(PODS '03). ACM, New York, NY, USA, 202-210. .

+

*Dinur, Irit and Kobbi Nissim. 2003. Revealing information while preserving privacy. In Proceedings of the twenty-second ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems(PODS '03). ACM, New York, NY, USA, 202-210. .

−

* Dwork, Cynthia, Frank McSherry, Kobbi Nissim, and Adam Smith. 2006. in Halevi, S. & Rabin, T. (Eds.) Calibrating Noise to Sensitivity in Private Data Analysis Theory of Cryptography: Third Theory of Cryptography Conference, TCC 2006, New York, NY, USA, March 4–7, 2006. Proceedings, Springer Berlin Heidelberg, 265-284, .

+

*Dwork, Cynthia, Frank McSherry, Kobbi Nissim, and Adam Smith. 2006. in Halevi, S. & Rabin, T. (Eds.) Calibrating Noise to Sensitivity in Private Data Analysis Theory of Cryptography: Third Theory of Cryptography Conference, TCC 2006, New York, NY, USA, March 4–7, 2006. Proceedings, Springer Berlin Heidelberg, 265-284, .

−

* Dwork, Cynthia. 2006. Differential Privacy, 33rd International Colloquium on Automata, Languages and Programming, part II (ICALP 2006), Springer Verlag, 4052, 1-12, .

+

*Dwork, Cynthia. 2006. Differential Privacy, 33rd International Colloquium on Automata, Languages and Programming, part II (ICALP 2006), Springer Verlag, 4052, 1-12, .

−

* Dwork, Cynthia and Aaron Roth. 2014. The Algorithmic Foundations of Differential Privacy. Foundations and Trends in Theoretical Computer Science. Vol. 9, Nos. 3–4. 211–407, .

+

*Dwork, Cynthia and Aaron Roth. 2014. The Algorithmic Foundations of Differential Privacy. Foundations and Trends in Theoretical Computer Science. Vol. 9, Nos. 3–4. 211–407, .

−

* Machanavajjhala, Ashwin, Daniel Kifer, John M. Abowd, Johannes Gehrke, and Lars Vilhuber. 2008. Privacy: Theory Meets Practice on the Map, International Conference on Data Engineering (ICDE) 2008: 277-286, .

+

*Machanavajjhala, Ashwin, Daniel Kifer, John M. Abowd, Johannes Gehrke, and Lars Vilhuber. 2008. Privacy: Theory Meets Practice on the Map, International Conference on Data Engineering (ICDE) 2008: 277-286, .

* Dwork, Cynthia and Moni Naor. 2010. On the Difficulties of Disclosure Prevention in Statistical Databases or The Case for Differential Privacy, Journal of Privacy and Confidentiality: Vol. 2: Iss. 1, Article 8. Available at: http://repository.cmu.edu/jpc/vol2/iss1/8.

−

* Kifer, Daniel and Ashwin Machanavajjhala. 2011. No free lunch in data privacy. In Proceedings of the 2011 ACM SIGMOD International Conference on Management of data (SIGMOD '11). ACM, New York, NY, USA, 193-204. .

+

*Kifer, Daniel and Ashwin Machanavajjhala. 2011. No free lunch in data privacy. In Proceedings of the 2011 ACM SIGMOD International Conference on Management of data (SIGMOD '11). ACM, New York, NY, USA, 193-204. .

−

* Erlingsson, Úlfar, Vasyl Pihur and Aleksandra Korolova. 2014. RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response. In Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security (CCS '14). ACM, New York, NY, USA, 1054-1067. .

+

*Erlingsson, Úlfar, Vasyl Pihur and Aleksandra Korolova. 2014. RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response. In Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security (CCS '14). ACM, New York, NY, USA, 1054-1067. .

−

* Abowd, John M. and Ian M. Schmutte. 2017 . Revisiting the economics of privacy: Population statistics and confidentiality protection as public goods. Labor Dynamics Institute, Cornell University, Labor Dynamics Institute, Cornell University, at https://digitalcommons.ilr.cornell.edu/ldi/37/

+

*Abowd, John M. and Ian M. Schmutte. 2017 . Revisiting the economics of privacy: Population statistics and confidentiality protection as public goods. Labor Dynamics Institute, Cornell University, Labor Dynamics Institute, Cornell University, at https://digitalcommons.ilr.cornell.edu/ldi/37/

−

* Abowd, John M. and Ian M. Schmutte. Forthcoming. An Economic Analysis of Privacy Protection and Statistical Accuracy as Social Choices. American Economic Review,

+

*Abowd, John M. and Ian M. Schmutte. Forthcoming. An Economic Analysis of Privacy Protection and Statistical Accuracy as Social Choices. American Economic Review,

−

* Apple, Inc. 2016. Apple previews iOS 10, the biggest iOS release ever. Press Release (June 13). https://www.apple.com/newsroom/2016/06/apple-previews-ios-10-biggest-ios-release-ever.html.

+

*Apple, Inc. 2016. Apple previews iOS 10, the biggest iOS release ever. Press Release (June 13). https://www.apple.com/newsroom/2016/06/apple-previews-ios-10-biggest-ios-release-ever.html.

* Ding, Bolin, Janardhan Kulkarni, and Sergey Yekhanin 2017. Collecting Telemetry Data Privately, NIPS 2017.

−

* http://www.win-vector.com/blog/2015/10/a-simpler-explanation-of-differential-privacy/

+

*http://www.win-vector.com/blog/2015/10/a-simpler-explanation-of-differential-privacy/

−

* Ryffel, Theo, Andrew Trask, et. al. "A generic framework for privacy preserving deep learning"

+

*Ryffel, Theo, Andrew Trask, et. al. "A generic framework for privacy preserving deep learning"

差分隐私上的阅读清单。2017.“当所有数据都是私人数据时，统计机构将如何运作?”。隐私与保密期刊7(3)。(幻灯片)

−

* “差分隐私: 非技术观众入门”，Kobbi Nissim，Thomas Steinke，Alexandra Wood，Micah Altman，Aaron Bembenek，Mark Bun，Marco gabordi，David r. o’brien，and Salil Vadhan，Harvard Privacy Tools Project，February 14,2018

+

*“差分隐私: 非技术观众入门”，Kobbi Nissim，Thomas Steinke，Alexandra Wood，Micah Altman，Aaron Bembenek，Mark Bun，Marco gabordi，David r. o’brien，and Salil Vadhan，Harvard Privacy Tools Project，February 14,2018

−

* Dinur，Irit and Kobbi Nissim。2003.在保护隐私的同时披露信息。在第二十二届 ACM SIGMOD-SIGACT-SIGART 数据库系统原理研讨会会议录(PODS’03)。ACM，纽约，纽约，美国，202-210. 。

+

*Dinur，Irit and Kobbi Nissim。2003.在保护隐私的同时披露信息。在第二十二届 ACM SIGMOD-SIGACT-SIGART 数据库系统原理研讨会会议录(PODS’03)。ACM，纽约，纽约，美国，202-210. 。

−

* Dwork、 Cynthia、 Frank McSherry、 Kobbi Nissim 和 Adam Smith。2006. in Halevi，s & Rabin，t.(Eds.)在密码学的私人数据分析理论中校准噪声的灵敏度: 第三次密码学理论会议，TCC 2006，纽约，纽约，美国，2006年3月4-7。美国国家科学院院刊，Springer Berlin Heidelberg，265-284，。

+

*Dwork、 Cynthia、 Frank McSherry、 Kobbi Nissim 和 Adam Smith。2006. in Halevi，s & Rabin，t.(Eds.)在密码学的私人数据分析理论中校准噪声的灵敏度: 第三次密码学理论会议，TCC 2006，纽约，纽约，美国，2006年3月4-7。美国国家科学院院刊，Springer Berlin Heidelberg，265-284，。

−

* 辛西娅。2006.差分隐私，第33届国际自动机，语言和编程学术讨论会，第二部分(ICALP 2006) ，Springer Verlag，4052,1-12，。

+

*辛西娅。2006.差分隐私，第33届国际自动机，语言和编程学术讨论会，第二部分(ICALP 2006) ，Springer Verlag，4052,1-12，。

−

* Dwork，Cynthia and Aaron Roth.2014.差分隐私的算法基础。理论计算机科学的基础与发展趋势。第一卷。9，Nos.3–4.211–407, .

+

*Dwork，Cynthia and Aaron Roth.2014.差分隐私的算法基础。理论计算机科学的基础与发展趋势。第一卷。9，Nos.3–4.211–407, .

−

* Machanavajjhala，Ashwin，Daniel Kifer，John m. Abowd，Johannes Gehrke，and Lars Vilhuber.2008.隐私权: 理论与实践的结合，国际数据工程会议2008:277-286，。

+

*Machanavajjhala，Ashwin，Daniel Kifer，John m. Abowd，Johannes Gehrke，and Lars Vilhuber.2008.隐私权: 理论与实践的结合，国际数据工程会议2008:277-286，。

−

* Dwork、 Cynthia 和 Moni Naor。2010.关于统计数据库中的披露防范的困难或者差分隐私的案例，隐私和保密期刊: 第一卷。2: Iss.1，第8条。网址: http://repository.cmu.edu/jpc/vol2/iss1/8。

+

*Dwork、 Cynthia 和 Moni Naor。2010.关于统计数据库中的披露防范的困难或者差分隐私的案例，隐私和保密期刊: 第一卷。2: Iss.1，第8条。网址: http://repository.cmu.edu/jpc/vol2/iss1/8。

−

* Kifer，Daniel and Ashwin Machanavajjhala.2011.数据隐私没有免费午餐。在2011年 ACM SIGMOD 国际数据管理会议记录(SIGMOD’11)。ACM，纽约，纽约，美国，193-204. 。

+

*Kifer，Daniel and Ashwin Machanavajjhala.2011.数据隐私没有免费午餐。在2011年 ACM SIGMOD 国际数据管理会议记录(SIGMOD’11)。ACM，纽约，纽约，美国，193-204. 。

−

* Erlingsson, Úlfar, Vasyl Pihur and Aleksandra Korolova.2014.RAPPOR: 随机可聚合隐私保护顺序响应。在2014年 ACM SIGSAC 计算机和通信安全会议(CCS’14)的会议记录中。ACM，纽约，纽约，美国，1054-1067。

+

*Erlingsson, Úlfar, Vasyl Pihur and Aleksandra Korolova.2014.RAPPOR: 随机可聚合隐私保护顺序响应。在2014年 ACM SIGSAC 计算机和通信安全会议(CCS’14)的会议记录中。ACM，纽约，纽约，美国，1054-1067。

−

* 以上，约翰 · m · 施穆特和伊恩 · m · 施穆特。2017 .重温隐私经济学: 人口统计和保密性保护作为公共产品。劳动动力学研究所，康奈尔大学，劳动动力学研究所，康奈尔大学， https://digitalcommons.ilr.cornell.edu/ldi/37/。即将到来。作为社会选择的隐私权保护与统计准确性的经济学分析。美国经济评论》，

+

*以上，约翰 · m · 施穆特和伊恩 · m · 施穆特。2017 .重温隐私经济学: 人口统计和保密性保护作为公共产品。劳动动力学研究所，康奈尔大学，劳动动力学研究所，康奈尔大学， https://digitalcommons.ilr.cornell.edu/ldi/37/。即将到来。作为社会选择的隐私权保护与统计准确性的经济学分析。美国经济评论》，

−

* 苹果公司，2016。苹果预览 iOS 10，史上最大的 iOS 发布。新闻稿(六月十三日)。Https://www.apple.com/newsroom/2016/06/apple-previews-ios-10-biggest-ios-release-ever.html.

+

*苹果公司，2016。苹果预览 iOS 10，史上最大的 iOS 发布。新闻稿(六月十三日)。Https://www.apple.com/newsroom/2016/06/apple-previews-ios-10-biggest-ios-release-ever.html.

−

* 丁、博林、贾纳丹•库尔卡尼及谢尔盖•叶卡宁二○一七。私下收集遥测数据 NIPS 2017。

+

*丁、博林、贾纳丹•库尔卡尼及谢尔盖•叶卡宁二○一七。私下收集遥测数据 NIPS 2017。

−

* http://www.win-vector.com/blog/2015/10/a-simpler-explanation-of-differential-privacy/

+

*http://www.win-vector.com/blog/2015/10/a-simpler-explanation-of-differential-privacy/

−

* Ryffel，Theo，Andrew Trask，et.艾尔。“一个保护隐私的通用深度学习框架”

+

*Ryffel，Theo，Andrew Trask，et.艾尔。“一个保护隐私的通用深度学习框架”

==External links==

−

* [https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/dwork.pdf Differential Privacy] by Cynthia Dwork, ICALP July 2006.

+

*[https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/dwork.pdf Differential Privacy] by Cynthia Dwork, ICALP July 2006.

−

* [http://www.cis.upenn.edu/~aaroth/Papers/privacybook.pdf The Algorithmic Foundations of Differential Privacy] by Cynthia Dwork and Aaron Roth, 2014.

+

*[http://www.cis.upenn.edu/~aaroth/Papers/privacybook.pdf The Algorithmic Foundations of Differential Privacy] by Cynthia Dwork and Aaron Roth, 2014.

−

* [http://research.microsoft.com/apps/pubs/default.aspx?id=74339 Differential Privacy: A Survey of Results] by Cynthia Dwork, Microsoft Research, April 2008

+

*[http://research.microsoft.com/apps/pubs/default.aspx?id=74339 Differential Privacy: A Survey of Results] by Cynthia Dwork, Microsoft Research, April 2008

−

* [http://video.ias.edu/csdm/dynamicdata Privacy of Dynamic Data: Continual Observation and Pan Privacy] by Moni Naor, Institute for Advanced Study, November 2009

+

*[http://video.ias.edu/csdm/dynamicdata Privacy of Dynamic Data: Continual Observation and Pan Privacy] by Moni Naor, Institute for Advanced Study, November 2009

−

* [http://simons.berkeley.edu/talks/katrina-ligett-2013-12-11 Tutorial on Differential Privacy] by [[Katrina Ligett]], California Institute of Technology, December 2013

+

*[http://simons.berkeley.edu/talks/katrina-ligett-2013-12-11 Tutorial on Differential Privacy] by [[Katrina Ligett]], California Institute of Technology, December 2013

−

* [http://www.cerias.purdue.edu/news_and_events/events/security_seminar/details/index/j9cvs3as2h1qds1jrdqfdc3hu8 A Practical Beginner's Guide To Differential Privacy] by Christine Task, Purdue University, April 2012

+

*[http://www.cerias.purdue.edu/news_and_events/events/security_seminar/details/index/j9cvs3as2h1qds1jrdqfdc3hu8 A Practical Beginner's Guide To Differential Privacy] by Christine Task, Purdue University, April 2012

−

* [https://commondataproject.org/blog/2011/04/27/the-cdp-private-map-maker-v0-2/ Private Map Maker v0.2] on the Common Data Project blog

+

*[https://commondataproject.org/blog/2011/04/27/the-cdp-private-map-maker-v0-2/ Private Map Maker v0.2] on the Common Data Project blog

−

* [https://research.googleblog.com/2014/10/learning-statistics-with-privacy-aided.html Learning Statistics with Privacy, aided by the Flip of a Coin] by Úlfar Erlingsson, Google Research Blog, October 2014

+

*[https://research.googleblog.com/2014/10/learning-statistics-with-privacy-aided.html Learning Statistics with Privacy, aided by the Flip of a Coin] by Úlfar Erlingsson, Google Research Blog, October 2014

*[https://www.belfercenter.org/publication/technology-factsheet-differential-privacy Technology Factsheet: Differential Privacy] by Raina Gandhi and Amritha Jayanti, Belfer Center for Science and International Affairs, Fall 2020

−

* Differential Privacy by Cynthia Dwork, ICALP July 2006.

+

* Differential Privacy by Cynthia Dwork, ICALP July 2006.

−

* The Algorithmic Foundations of Differential Privacy by Cynthia Dwork and Aaron Roth, 2014.

+

*The Algorithmic Foundations of Differential Privacy by Cynthia Dwork and Aaron Roth, 2014.

−

* Differential Privacy: A Survey of Results by Cynthia Dwork, Microsoft Research, April 2008

+

*Differential Privacy: A Survey of Results by Cynthia Dwork, Microsoft Research, April 2008

−

* Privacy of Dynamic Data: Continual Observation and Pan Privacy by Moni Naor, Institute for Advanced Study, November 2009

+

*Privacy of Dynamic Data: Continual Observation and Pan Privacy by Moni Naor, Institute for Advanced Study, November 2009

−

* Tutorial on Differential Privacy by Katrina Ligett, California Institute of Technology, December 2013

+

*Tutorial on Differential Privacy by Katrina Ligett, California Institute of Technology, December 2013

−

* A Practical Beginner's Guide To Differential Privacy by Christine Task, Purdue University, April 2012

+

*A Practical Beginner's Guide To Differential Privacy by Christine Task, Purdue University, April 2012

−

* Private Map Maker v0.2 on the Common Data Project blog

+

*Private Map Maker v0.2 on the Common Data Project blog

−

* Learning Statistics with Privacy, aided by the Flip of a Coin by Úlfar Erlingsson, Google Research Blog, October 2014

+

*Learning Statistics with Privacy, aided by the Flip of a Coin by Úlfar Erlingsson, Google Research Blog, October 2014

*Technology Factsheet: Differential Privacy by Raina Gandhi and Amritha Jayanti, Belfer Center for Science and International Affairs, Fall 2020

差分隐私: Cynthia Dwork，ICALP July 2006。差分隐私的算法基础》，Cynthia Dwork 和 Aaron Roth，2014年。2013年12月，加州理工学院卡特里娜 · 利格特教授，差分隐私，差分隐私，差分隐私实用指南，克里斯汀 · 特拉克，普渡大学,2012年4月

−

* 私人地图制作者 v0.2 on the Common Data Project Blog

+

* 私人地图制作者 v0.2 on the Common Data Project Blog

−

* Learning Statistics with Privacy，added by the Flip of a Coin by úlfar Erlingsson，Google Research Blog，October 2014

+

*Learning Statistics with Privacy，added by the Flip of a Coin by úlfar Erlingsson，Google Research Blog，October 2014

−

* Technology Factsheet: 差分隐私地图制作者 Raina Gandhi and Amritha Jayanti，Belfer Center for Science and International Affairs，Fall 2020

+

*Technology Factsheet: 差分隐私地图制作者 Raina Gandhi and Amritha Jayanti，Belfer Center for Science and International Affairs，Fall 2020

−

[[~~Category~~:Differential privacy| ]]

+

[[index.php?title=分类:Differential privacy| ]]

[[Category:Theory of cryptography]]

[[Category:Information privacy]]

Litinunispazio97

23

个编辑