"Multilingual Text Analysis on On-Line Text for Early Warning" Generally, on-line blog discussion of a topic, even heated discussion, stays on-line. Sometimes, however, that discussion spills over into concrete, real-world consequences. Sandia Laboratories, as part of its Networks Grand Challenge research project (ngc.sandia.gov), has derived methods to provide early warning of such spill-over on topics of known interest. We will describe these methods, and how they combine analysis techniques from the diverse domains of text, network, and social dynamics. Further, we will show how to extend them to multilingual discussion in a translation-free fashion, leading to multilingual sentiment analysis methods that can provide early warning of topics to watch out for. Bio: Philip Kegelmeyer (EE Ph.D, Stanford) is a Senior Scientist at Sandia National Laboratories in Livermore, CA. He has twenty years experience inventing, tinkering with, and quantitatively improving machine learning algorithms. His work has resulted in over fifty refereed publications, two patents, and commercial software licenses. He recently served as the Principle Investigator for a large (three years, forty part-time staff) "Grand Challenge" internal research and development project devoted to network analysis. His current research addresses ensemble methods for unsupervised learning in graph data and supervised learning in text.