A time to remember, a time to forget: Enabling people to control long term sensor data
Access status:
USyd Access
Type
ThesisThesis type
Doctor of PhilosophyAuthor/s
Barua, DebjaneeAbstract
The key contribution of this thesis is to propose a conceptual model and then demonstrate its use in designing the infrastructure and user interfaces for user modelling where users can control the remembering and forgetting of ubicomp sensor data in their long term user models. ...
See moreThe key contribution of this thesis is to propose a conceptual model and then demonstrate its use in designing the infrastructure and user interfaces for user modelling where users can control the remembering and forgetting of ubicomp sensor data in their long term user models. This is the first research work that has explored forgetting mechanisms and developed foundations for user control over them in the long term user models that deal with sensor data. Emerging technologies in ubiquitous computing are creating the infrastructure for unobtrusive sensing to create large stores of personal data. This data has immense potential to serve many important roles in people’s lives (e.g., long term wellbeing goals). Data captured by these trackers can be stored in a long term user model to support personalised wellness applications. One of the challenges of such model is that it accumulates huge amount of sensor data over time. We argue that this creates the need for forgetting mechanisms. Drawing inspiration from forgetting in human memory and computing systems, we defined “Mneme” conceptual model. Mneme’s infrastructure included four levels of user model stores (Active, Long term, Archive and Trash stores) and forgetting mechanisms (e.g., decay, deletion, and archival). Our insight was to frame the user control interfaces around people’s long term goals. We conducted a questionnaire study with 107 participants that helped us gain insights into people’s perceptions of management of ubicomp sensor in a user model. To validate our conceptual model for forgetting, we implemented it in an infrastructure with Mneme user control interface. We assessed the user interface in two user studies: an in-lab Think Aloud study with 17 participants and a Field Study with 14 participants over a period of 12 weeks. The results indicate that users could perform the core management tasks without training; they understood the key concepts of Mneme and considered the mechanisms to be useful.
See less
See moreThe key contribution of this thesis is to propose a conceptual model and then demonstrate its use in designing the infrastructure and user interfaces for user modelling where users can control the remembering and forgetting of ubicomp sensor data in their long term user models. This is the first research work that has explored forgetting mechanisms and developed foundations for user control over them in the long term user models that deal with sensor data. Emerging technologies in ubiquitous computing are creating the infrastructure for unobtrusive sensing to create large stores of personal data. This data has immense potential to serve many important roles in people’s lives (e.g., long term wellbeing goals). Data captured by these trackers can be stored in a long term user model to support personalised wellness applications. One of the challenges of such model is that it accumulates huge amount of sensor data over time. We argue that this creates the need for forgetting mechanisms. Drawing inspiration from forgetting in human memory and computing systems, we defined “Mneme” conceptual model. Mneme’s infrastructure included four levels of user model stores (Active, Long term, Archive and Trash stores) and forgetting mechanisms (e.g., decay, deletion, and archival). Our insight was to frame the user control interfaces around people’s long term goals. We conducted a questionnaire study with 107 participants that helped us gain insights into people’s perceptions of management of ubicomp sensor in a user model. To validate our conceptual model for forgetting, we implemented it in an infrastructure with Mneme user control interface. We assessed the user interface in two user studies: an in-lab Think Aloud study with 17 participants and a Field Study with 14 participants over a period of 12 weeks. The results indicate that users could perform the core management tasks without training; they understood the key concepts of Mneme and considered the mechanisms to be useful.
See less
Date
2016-03-21Licence
The author retains copyright of this thesis. It may only be used for the purposes of research and study. It must not be used for any other purposes and may not be transmitted or shared with others without prior permission.Faculty/School
Faculty of Engineering and Information Technologies, School of Information TechnologiesDepartment, Discipline or Centre
Computer Human Adapted Interaction Research GroupAwarding institution
The University of SydneyShare