|
||||
|
Situational Awareness
Self-awareness in a content-centric Internet | Permalink In recent years, user-generated content and participatory web-based services have become more popular on the Internet. Examples include ‘traditional’ formats, such as photographs (Flickr) and blogs (livejournal.com, blogger.com), social networks (Facebook, LinkedIn), and micro-blogging (Tumblr, Twitter). More recently, location-based services (foursquare.com) and ad-hoc sharing (color.com) have begun to gain popularity. Often grouped together under the term Web 2.0, much of this content relates to places and activities. The diversity of content has led to the 'long-tail' effect, in which a large amount of content is only relevant to a small and very specific audience. Accessing such content through search engines can be difficult, as it can easily be buried among material that is more popular to a wider audience. This is particularly the case in spatial and temporal scenarios, such as finding content on the move, while away from the desktop computer. Discoverability is also an issue for content within a search-based paradigm, as users may not even know that interesting and relevant material exists. In our work, we are looking for new ways to autonomously provide and manage content relevant to the individual user. The project is inspired by the cognitive processes that humans exhibit for self-awareness. Examples include the heuristics and cues that we subconsciously use every day for rapid decision making and negotiating conflicting signals in physical and social situations. In our project, our system makes information available using self-awareness that is based on knowledge of how information on the Internet is consumed, provided, and managed (see Figure 1). Consumption describes how users consume and interact with content (for example, how they move from one item to the next). Provision refers to mobility, in which the system pushes spatially and temporally relevant content to users as they move from place to place in a city. Finally, effective management allows content to be responsive to the needs of the content provider and the network by managing the accessibility and availability of information. Figure 1. Self-awareness based on the way in which information is consumed, provided, and managed on the Internet. 
To apply these three scenarios, we need to embed self-awareness at the user, content and network levels. We aim to achieve this using a tripartite model for cognition. The three components of this model are acquiring knowledge, decision making and learning. Acquiring knowledge refers to simple, predefined processes that provide cues and monitor the external environment. In decision making, more complex processes operate with partial or aggregated information to make choices and establish priorities. Learning refers to longer-term activities that operate at a high level, including memory, to further understand and adapt both knowledge acquisition and decision making. In each of the components of the model, there is a human dimension that dictates how individuals go about the related functions and prioritize different issues. These dimensions relate to psychological and human traits (such as personality, mobility behaviour and different cognitive preferences) and demonstrate a further way in which individual relevance and awareness can be customized. We are developing computational processes that support this tripartite model. For different scenarios, these processes become nodes within the system. The nodes are intelligent software agents that allow users to gain awareness within a content-centric Internet. To demonstrate the robustness of the approach, we are tackling some diverse deployments of the nodes. These include providing relevant content to mobile users' smart phones (for example, identifying diverse, but relevant content about places and affordances when walking through a city), enabling opportunistic networking of devices to push content of interest in a resource-efficient manner, providing content that supports real-time decision making, personalizing content streaming and controlling access to multimedia content. To illustrate the approach, consider the problem of pushing relevant content to individuals as they move through their environment. A diverse range of cues can suggest contextual information that may be useful in determining the relevance of information. For example, we could use people's history of content consumption and creation to determine interest profiles to prioritize sharing.1 The system can classify and analyse the locations a person visits, and feed this information to the spatial domain,2 while past social interactions can predict future communities.3 Furthermore, recent work4 has shown how personality traits are linked to media preferences. One of the challenges addressed in our work is how best to collect, combine, and use such cues to efficiently filter and forward content in real-time. From this approach, we are gaining important insights into human behaviour and how it is manifested in technology. These include understanding human mobility and personality as well as the effect of individuals' need for cognition. More generally, the project is providing a new approach to socio-technical system design that is flexible, behaviourally based and considerate of deeper personal traits. Future work will focus on developing a flexible and reusable self-aware node built upon the tripartite cognitive model, and demonstrating its adoption in diverse content-centric scenarios. References
Stay Informed
|
||||
|
||||