 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
 |
|
|||||||||||||
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|
|||
 |
 |
 |
 |
|
||||||||||
 |
|
|||||||||||||
 |
|
|||||||||||||
 |
 |
|
||||||||||||
| Seminar Schedule (Seminar Organizer: Prof. Ziv Bar-Joseph) | |
Machine Learning Lunch Seminar
|
Spring 2008 Seminar
Date: January 28, 2008 Time: 4:30 PM - 5:30 PM (Refreshments at 4:15) Location: 1305 Newell-Simon Hall Speaker: Lucio Soibelman Associate Professor, Carnegie Mellon University Title: Data Management and Knowledge Discovery for Advanced Infrastructure Systems Abstract: It is certainly no surprise that construction and operations of infrastructure systems require a huge amount of information from specifications, plans, construction documents, inventory management, cost estimating, and scheduling, for the construction phase and maintenance records, and inspections data from the operations phase. As this industry adopts new computer technologies, computerized construction/operations data are becoming more and more available. There exist numerous opportunities to exploit and extract knowledge from the vast amount of infrastructure data. Unlike much previous research in Knowledge Discovery in Databases (KDD) that has been successfully applied in several domains, in the infrastructure domain, however, the data are of multiple types and from many different sources, some with very low quality. Professor Soibelman will be presenting his research on the organization of such diverse data into forms that are amenable to the application of knowledge discovery methods. He will be introducing the results obtained from several studies developed by his research group. The first research to be presented is a project that developed a framework for construction knowledge discovery with the objective of providing a comprehensive data mining methodology to be applied in other construction management areas. The outcome of this research was a methodology that permits construction project managers who have no DM/KDD expertise to discover knowledge from their own project databases. The second research to be presented is a study that developed methods for integrating unstructured text documents and pictures in A/E/C model based systems. In this study, automated processes for retrieval, classification, and integration of unstructured documents in A/E/C model based systems were explored. Specifically, a combination of techniques from the areas of information retrieval, text mining, image reasoning, and computer vision were analyzed to develop intelligent search engines to identify documents/pictures relevant to each component of the project model. Finally Professor Soibelman will introduce his current research activities among them the development of image reasoning methodologies with the objective of supporting a sewer inspection robot to detect and automatically classify sewage defects, a geospatial data warehouse decision support framework for electricity production vulnerability assessment to help decision makers understand the impact of fuel delivery disruption and the vulnerabilities in the coal transportation system, and an electricity metering data analysis tool to support better user decision making for home energy consumption.
