Information Analytics Lab
Goal
Enable near real-time business intelligence with robust, scalable data management, data-intensive analytics and fusion of structured and unstructured information. Fulfill the demand for timely delivery to enterprise customers of information that is described in the language of businesses, processed within their business context, and automatically selected, gathered and transformed from diverse, heterogeneous, distributed sources.
Impact
Information is a critical enterprise asset displaying exponential growth. Information can be analyzed for decisions and new insights, and turned into knowledge delivered into user contexts. It powers modern businesses and scientific research and is the lifeblood of enterprise business processes.
Research in information analytics builds on today’s business intelligence and related information-management technology foundation, and dramatically expands its horizon to achieve near real-time data capture and integration, adaptive and closed-loop operational business intelligence supporting a wide user base, data-intensive parallel analytics, and fusion of business intelligence with search and information extraction over structured and unstructured information from a wide variety of internal and external sources and data feeds.
The research draws requirements from emerging applications ranging from consumer industries delivering better customized products and services, supply chains synchronizing with real-time demand signals and scientists improving their ability to access and analyze massive amounts of data.
End users will be able to work more flexibly with vast amounts of complex information, by using complete, timely, contextually minimized information, which will be integrated, optimized, and personalized using innovative context-modeling and content-mining algorithms applied to rich data complying with XML, and related open standards.
Research
- real-time continuous feed of business events into business intelligence systems and stream data processing to reduce the latency between the time an event occurs and the time quality data is available for decision-making
- workload management techniques to enable self-managing, adaptive tuning for complex, mixed workloads, and robust query processing, storage and access methods that adapt gracefully to changing workload
- algorithms for scalable data-intensive analytics, including real-time adaptation of models and extraction of information from unstructured data
- fusion of search, extraction, query and analytics to enable integration of structured and unstructured information in solution contexts
- use of parallel architecture on multi-core clusters and emerging memory hierarchy and storage devices in large-scale data management, and use of distributed infrastructure to further scale across geographical and organizational boundaries
- emerging, innovative vertical applications enabled by the next generation information platform
- dramatically advance the state of the art in rich-information sourcing
- deliver breakthroughs in modeling information-consumption context
- contextually optimize information quality and use of human attention
- allow users to impose their personal view on Web and corporate information
Director: Umeshwar Dayal