Using Intelligent Agents for Social Sensing across Disadvantaged Networks

Users who retrieve information across disadvantaged networks need to do so in such a way as to minimize network performance impact while maximizing the usefulness and quality of information (QOI) received. Taking advantage of features from all three network genres (telecommunication, information, and social) will enable this balancing act. These users also need to interact using unstructured, ad-hoc queries to decrease the cognitive overload of specialized training or the necessity of learning a constrained language. High QOI can be maintained if an intelligent agent on the network can use social sensing to capture the intent of the query and identify the implied task. Knowing the task will allow other agents that service the requests to filter, summarize, or transcode data prior to responding, lessening the network footprint. This paper describes an approach that uses natural language processing (NLP) techniques, multi-valued logic based inferencing, network status checking, and task-relevant metrics for information retrieval. This research effort has resulted in a low-level NLP approach that can be used to capture intent from unstructured text, a quality metric formed from intrinsic and extrinsic attributes of the objects and the tasks under consideration, and a simple inferencing approach to allow intelligent agents to make quality assessments, delivering the appropriate form of the information that will lessen the impact on a disadvantaged network.