Privacy in Context

隐私是关于选择和背景的。选择至关重要，因为每个人都必须能够定义他们想要持有的东西是私人的。上下文定义了共享私人信息所围绕的框架以及共享后对其使用的限制。每个人将对什么是私人做出不同的选择，并将有多个应在其中共享私人信息的上下文。需要隐私保护和感知数据系统，该系统尊重个人选择并允许用户定义由于不同上下文而引起的差异。数据隐私必须以类似的方式定义。任何关于我们收集的数据都应该按照个人对数据的偏好来处理(或者更理想地说，更广泛地说，是关于谁收集了数据)。这些首选项定义了提供者对数据项的敏感度的选择，以及何时和谁可以共享数据(上下文)。 目前的隐私方法要么对收集的任何数据强加一个通用的隐私结构，充其量只提供一个选择加入或选择退出的参与模式。我们的方法认为，只有在用户允许作为个人选择的情况下，数据才应该包括在分析活动中。我们正在调查与连接环境和数据管理方面的选择相关的端到端挑战。我们确定了三个主要的新贡献：首先，我们将开发一个隐私承诺模型，该模型将捕获数据隐私的关键方面，包括隐私元数据，我们将开发一个新的隐私本体，明确地捕获允许基于上下文的不同用途的个人偏好。其次，我们将开发有效地将收集的数据附加到基于知情同意的隐私承诺的技术。数据和隐私承诺必须不可撤销地附加在一起。这将需要新的数据处理范例和数据结构，能够在保持这种紧密联系的同时，实现有效的数据处理和分析。一个成功的战略还将促进更好的分析，因为它提高了数据完整性，从而增加了效用。第三，我们将开发和/或调整系统，以高效地处理、存储和通信数据，并以永远不会破坏联系的方式，从而确保在过程的每一步都履行隐私承诺。我们将研究一些不同的数据收集范例来测试我们的隐私模型，包括传感器系统(例如。物联网)、网络浏览活动、销售点系统、智能公用事业以及市场分析系统等大数据系统。 加拿大人高度重视隐私，因为它是现代民主的基石。隐私保护在过去20年里慢慢退化，但它的重要性现在被视为一个更优先的问题。我的工作将开发保护这一价值的系统，并通过允许个人控制他们的哪些信息应该被保密来做到这一点。

Privacy is about choice and context. Choice is critical because each individual must be able to define what they want to hold as private. Context defines the framework around which private information is shared and the limits on its use once it has been shared. Each individual will make different choices about what is private and will have multiple contexts in which the private information should be shared. Privacy preserving and aware data systems are needed that respect individual choice and allow the user to define variance arising as a result of different contexts. Data privacy must be defined analogously. Any data collected about us should be treated in conformance with individual preferences about the data (or ideally, and more generally, about whom the data is collected.) These preferences define the provider's choices about a data item's sensitivity, and when and with whom the data can be shared (context). Current privacy approaches either impose a generic privacy structure on any data collected and at best provide only an opt-in or opt-out participation model. Our approach argues that data should only be included in analyses activities if the user permits it as a personal choice. We are investigating end-to-end challenges associated with connecting context and choice with respect to data management. We identify three major novel contributions: First, we will develop a privacy commitment model that captures key aspects of data privacy including privacy meta-data and we will develop a new privacy ontology that explicitly captures individual preferences that allow for different uses based on context. Secondly, we will develop techniques that efficiently attach collected data to privacy commitments based on informed consent. The data and privacy commitments must be irrevocably attached. This will require new data processing paradigms and data structures capable of maintaining this tight linkage while permitting efficient data processing and analytics. A successful strategy will also facilitate better analytics because of increased data integrity thereby increasing utility. Thirdly, we will develop and/or adapt systems that process, store, and communicate data efficiently and in such a way that the linkage is never broken thereby guaranteeing the privacy commitments are honoured at each step of the process. We will investigate a number of different data collection paradigms to test our privacy model including sensor system (eg. IoT), web browsing activities, point-of-sales systems, smart utilities, and “big data” systems such as market analysis systems. Canadians highly value privacy as it is the cornerstone of modern democracies. Privacy protection has slowly degraded over the past two decades but its importance is now seen as a much higher priority issue. My work will develop systems that protect this value and do so by allowing individuals to control what about them should be held as private.