Current ResearchOngoing research is organized around two key themes:
Inferring Knowledge from Data:Analyze large archives of real-world observations to learn about correlations, trends, and patterns in business problems, customer behavior, and markets.
Media Analytics: extract knowledge of patterns, events, and anomalies from multi-media signal streams including text, audio, video, and sensor output; for intelligent coding, routing, archiving, and retrieval. We collaborate with the Multimedia Research Center on efforts like video analysis.
Recommender Systems and Service Personalization: infer and summarize succinctly user preferences and habits in order to personalize network services and applications.
Services Data Mining: from large databases of technical support service transactions, find drivers of customer satisfaction, distributions of service cost and equipment outage, and characterize patterns of usage of knowledge base documents.
Real-Time Monitoring / Prediction:Enhance communication systems and services with intelligence by enabling them to discover and respond to anomalies.
Smart Grid Monitoring and Data Mining: short-term load forecasting, detection of grid and customer anomalies.
Internet Traffic Analysis: find intrusions, malfunctions, congestions in networks by real-time analysis of traffic traces.
Wireless Indoor Tracking: locate people and assets in an indoor environment through radio frequency communications.
Acoustic Monitoring: determine stage of life of equipment cooling fans by analyzing characteristics of mechanical noise.
System Diagnostics: develop model-based or data-driven diagnostic tools to uncover performance anomalies their root causes in optical or radio network equipment systems.
More about our research activities are described in:
Opportunities with Automated Inference with Data in Tomorrow's Communication Networks and Services, by Tin Kam Ho, Thomas Bengtsson, Bell Labs Technical Journal, 2009.