SRIs Machine Reading Program is creating an automated reading system that makes the information in natural language texts accessible to a range of formal reasoning systems.
In the FAUST project, SRI proposes an architecture based on statistical joint inference over probabilistic relational models. Supporting the simultaneous consideration of random variables enables the leveraging of all mutually constraining information and the integration of information across sentences and texts. This joint inference engine will integrate information from more specialized inference modules, in particular, an ensemble of natural language modules. Such integration requires the coordination or alignment of representations from multiple levels of analysis and multiple texts.
SRI is continually improving these alignments using machine learning techniques. This approach builds on significant recent advances in probabilistic representation and joint inference, and will enable the system to consider the widest range of both linguistic and extra-linguistic evidence, using the same mechanisms that are used for the integration of linguistic information across levels.
SRIs subcontractors are Columbia University, Stanford, University of Massachusetts, University of Illinois, University of Washington, University of Wisconsin, Stanford CSLI, and Wake Forest University.