Context-Aware Computing: Beyond Search and Location-Based Services
This month's theme includes the following featured articles:
The Telecom Industry and Context Awareness
Location-based services provide a variety of revenue-generating possibilities for telecom companies. Take a look at some proposed business roles that would leverage context awareness. More »
A Contextualized and Personalized Approach for Mobile Search
Learn about personalization strategies that aim to improve mobile search by explicitly and implicitly inferring context from the user's environment. More »
Timely and Keyword-Based Dynamic Content Selection for Public Displays
A recommender system for public displays can autonomously select relevant content from Internet sources using a keyword-based place model as input. More »
Context-Aware Automatic Query Refinement Using Indian-Logic Based Ontology
Query refinement can improve search results, but most existing approaches don't consider context. This novel system identifies the user's context of search and uses context ontologies to make search results more relevant. More »
Defeasible Contextual Reasoning with Arguments in Ambient Intelligence
Contextual reasoning in ambient intelligence environments is a very challenging task. This proposed solution is based on the Multi-Context Systems paradigm. More »
Context is the unstated actor in human communications, actions, and situations. It makes our communications efficient, our commands actionable, and our situations understandable by those—devices, people, or organizations—seeking to provide us with content or services. The advent of context-aware computing is therefore concomitant with the increased embedding of technology into our personal and social environments.
Why Context-Aware Computing?
Proactive, contextualized delivery of information, alerts, and advertisements (including content and product recommendations) represents significant commercial opportunities for all types of service providers. It stands to drive the user experience of mobile consumers well beyond having to search for information and beyond location-based services that adapt to mere vestiges of our larger context.
Context mediation also enables orchestration of Web services, thereby yielding actionable interpretation of spoken high-level commands, a key technology in creating truly smart personal assistants. It stands to deliver novel conversational affordances into interfaces ranging from devices to public display kiosks.
Comprehensive capture, representation, communication, gathering, and brokering of the larger user context, according to Gartner (A. Johnson, J. O'Brien, and G. Alvarez, Context Shapes Demand at Moments of Truth, Industry Research report ID: G00200528, Gartner, Inc., 25 May 2010), represents billions of dollars per year of revenue opportunity in industries such as travel and retail. Context awareness sharpens relevance when responding to user-initiated actions (such as product search and support calls). It also enables proactive communications through pattern analytics over both user behavior and user environment.
Physical or sensed context is captured using GPS and other devices. It is decorated using linked open data, other aspects of user-provided context, and social networks, and further analyzed, in complex commercial ecosystems of context providers and brokers. This derived context becomes the target of content providers and advertisers, who receive context views of users created using context caching, indexing, and partitioning algorithms. The content and services can either be pushed into context by the provider or pulled into it by the user.
Furthermore, context is itself a type of data for use by various operations, such as resolving ambiguous references, resolving inconsistent facts, and automated reasoning. For example, the words "the Sienna" refer to a specific entity when used in my home context, but to an entire family of cars when used in the automobile industry context. And, in pulling together user reviews of its features, the Sienna might be considered good in one user's context and bad in another user's context. Characterizing and clustering such contexts of mention saves us from inappropriate use of averaging across mutually inconsistent data sources.
The real power of context, and so far the area of its greatest unrealized potential, is in automated reasoning. Context can be viewed as a theory—a collection of facts (entities and their inter-relationships) and rules—describing the environment of a user or an event. By lifting these facts and rules into a particular (sparsely described) situation, context makes possible inference. Inference is the foundation that makes devices and services look smart, and capable of working with mere clues as to the user's intentions and preferences. Contextual inference will ultimately give us the technology for making even better recommendations, be they songs, product suggestions, or other search results.
Selected Articles for Context-Aware Computing
In 2007, Jordi Docter, Carlo Alberto Licciardi, and Marco Marchetti surveyed the state of context-aware computing in "The Telecom Industry and Context Awareness." They laid down the blueprint for an industry-wide ecosystem of context providers and brokers, enmeshed with content providers and service providers. In 2010, much of what they lamented—such as the lack of "successful implemented context aware services in the market"—has started to change, but, interestingly, along the lines they had suggested in this paper.
In "Timely and Keyword-Based Dynamic Content Selection for Public Displays," Fernando Reinaldo Ribeiro and Rui JosÃ© use the approach of creating a shared context across users through place-centric content integration and summarization, rather than direct modeling and sensing of diverse user contexts. Even their simplistic models of context illustrate the use of context in topic disambiguation.
In "A Contextualized and Personalized Approach for Mobile Search," Feng Gui, Malek Adjouadi, and Naphtali Rishe present device-based mechanisms of context capture and a server-side mechanism for context modeling and analysis that supports contextualized expansion of search queries and contextualized ranking of search results. Their work further illustrates the role of context in resolving ambiguous query keywords.
S.M. BaalaMithra and S.M. SominMithraa describe more explicit modeling of the context of a search user, albeit on a per-session basis, in "Context-Aware Automatic Query Refinement Using Indian-Logic Based Ontology." Their CAQR algorithm models the various contexts of ambiguous concepts in a context ontology—for instance, explicitly representing taxonomical and other relationships describing the five different senses of the word "apple." Two other uses of context are well illustrated by CAQR: first, clustering of search results by context, and second, ranking of contexts and re-ranking of search results based on the user's selection from a ranked listing of contextually clustered results. The paper shows massive improvements in precision and contextualization of search using this algorithm.
Antonis Bikakis and Grigoris Antoniou present an algorithm for the challenging task of integrating possibly inconsistent information from multiple partial contexts accumulating in a distributed fashion. Their article, "Defeasible Contextual Reasoning with Arguments in Ambient Intelligence," also illustrates the representation of context as a theory and the use of advanced logics and logic-based systems in computing with context.
The blueprint for an ecosystem of context providers and brokers is falling into place. Starting with contextualized delivery of information, alerts, and advertisements, the opportunities for context-aware computing are as abundant as the places in which technology will become embedded in our lives. By building sensors for gathering context clues and platforms for integrating and serving context, we as engineers can accelerate the development of context-aware applications. In March 2012, IEEE Internet Computing magazine plans to run a special issue on the same topic (Beyond Search), in which we will share further progress on this important theme.
We hope you enjoy reading these articles as much as we did. For further information, take a look at these related links.
Managing Healthcare through Social Networks
Surveys show an increased reliance on physician and patient social networks, which promise to transform healthcare management. But challenges such as privacy and data accuracy remain. More »
What's Wrong with Access Control in the Real World?
Effective security requires looking at an entire system—looking at only one piece leads to security trouble. This extends to looking at only what we pretend a system does, rather than what it actually does in the real world. More »
The Future of Architectural Simulation
Intel's Joel Emer and Microsoft Research's Doug Burger spearhead a debate on the future of architectural simulation, captured from the 2009 Computer Architecture Research Directions workshop. More »
Toward Business Transaction Management in Smart Service Networks
A novel business transaction model for smarter service networks leverages pooling real-time information from software services and sensor networks to more effectively manage fine-grained tenets of service-based applications. More »
Optical Interconnect for High-End Computer Systems
Large-scale chip arrays can use high-bandwidth optical links for communication that enables significant levels of device integration. More »
Preventing Future Oil Spills with Software-Based Event Detection
Complex event processing systems detect problems in mission-critical, real-time applications and generate intelligent decisions to modulate the system environment. More »
The Top 10 Burning Research Questions from Practitioners
Does academic research meet software practitioners' requirements or expectations? The answer is often no. XP 2010 attendees registered their top 10 requests. Add your voice. More »
Real-Time Environmental Monitoring and Notification for Public Safety
A wireless sensor-network database application combines environmental monitoring and Web-based, real-time data reporting and visualization for both real-time and archived data. More »
Visualization in Transportation: Challenges and Opportunities for Everyone
Transportation is the backbone of civilization and the reason for our economic prosperity. This article introduces visualization research in the transportation community, discusses research opportunities, and encourages the computer graphics community to get involved. More »
Parallel Traffic Management for the 2010 Asian Games
The Asian Games in Guangzhou involves 42 sports, 100,000+ people, 58 existing facilities, and 12 new stadiums, so transportation management must be safe and effective. More »