Welcome Silvia Ferrari
A Professor in the Sibley School of Mechanical and Aerospace Engineering, Silvia Ferrari joined Cornell this summer and is currently Director of the Laboratory for Intelligent Systems and Controls (LISC). An aerospace engineer by training, Ferrari develops new methods rooted in machine learning and systems theory to design intelligent autonomous systems that are able to learn and discover new information over time. Her contributions include the development of new theories and algorithms on the learning and approximation properties of graphical models, such as neural and probabilistic networks, as well their applications in many areas of science and engineering, such as reconfi gurable aircraft control and robotics. Professor Ferrari developed new methods for adaptive dynamic programming, reinforcement learning, optimal control, and information-driven planning and control for distributed systems and mobile sensor networks. Recent contributions also include the development of new mathematical models of learning and plasticity uncovered from biological brains, as well as cognitive models of complex decision-making derived from data.
Prior to joining Cornell, Ferrari was professor of engineering and computer science at Duke University and a faculty member of the Duke Institute for Brain Sciences (DIBS). While at Duke, she founded the Integrative Graduate Education and Research Training (IGERT) Program on Wireless Intelligent Sensor Networks (WISeNet), which now includes over 35 faculty members housed in 12 entities including Engineering, the School of the Environment, Arts and Sciences, and the Environmental Protection Agency (EPA), the Naval Undersea Warfare Center (NUWC), and companies such as Ferrari S.p.A. and BAE Systems. The purpose of this program is to prepare Ph.D. students for interdisciplinary research in wireless sensor networks that process, store, and learn from data so as to improve their ability to gather information over time. By participating in field experiments led by EPFL, University of Ulster, University of Cagliari, and University of Bologna, the students in this program have been able to contribute first hand to unprecedented observations of environmental and ecological processes, and to develop methods for more effective and reliable use of sensors for defense and national security. Ferrari’s collaborations and partnerships in these areas continue here at Cornell, where LISC students and researchers conduct collaborative research and field experiments aimed at demonstrating information-driven planning and control methods for unmanned vehicles engaged in underwater demining, mobile methane sensing, search and rescue, and surveillance.
Although sensor technologies vary considerably, from embedded sensor systems on unmanned vehicles to tiny “smart dust” sensors gathering and relaying environmental data to a central computer, a unifying paradigm that has recently emerged in the literature is to treat the sensor network as a system of dynamic information-gathering agents whose performance depends on the environment they are deployed to observe. As a result, the same environmental models that require the use of sensor data to update environmental predictions and forecasts can be used to manage and control the sensors intelligently over time. Professor Ferrari and her group develop theory and computational methods to allow wireless sensor networks to bett er process and learn from sensed information, adapt autonomously to unanticipated situations, self-coordinate to meet multiple objectives and constraints, and evolve over time to exhibit greater functionality in changing and complex environments. An intelligent system is able to process, store, and retrieve information, identify changes, plan actions, learn from past experience, and adjust to new environments and situations to improve its performance over time. A fundamental feature of the brain is that it transforms sensory inputs into appropriate motor outputs, and when the sensor inputs change, the motor outputs can adapt. Adaptation and learning are central properties of these sensori-motor transformations which must continuously face new circumstances, from growth and development to migratory or seasonal change, and adapt to new complex and unstructured environments. Professor Ferrari’s research aims at reproducing in artifi cial systems some of the capabilities of biological sensori-motor systems such as coordinating sensor movements and fusing heterogeneous data, while performing complex motor tasks such as searching for hidden clues or targets, landing on a moving surface, or moving across an obstacle-populated room.
Ferrari’s work has been recognized through many awards and honors, including the Presidential Early Career Award for Scientists and Engineers (PECASE), the NSF CAREER Award, the ONR Young Investigator Award, and the International Crime Analysis Association Research Award. Ferrari is an IEEE Senior Member and Chair of the IEEE Technical Committ ee on Adaptive Dynamic Programming and Reinforcement Learning. She also serves on the Advisory Board of several international institutes and multi-institution projects, such as the Intelligent Systems Research Center (ISRC), University of Ulster, UK, and the SHERPA European Union (EU) Project. Ferrari has co-authored over 90 refereed publications spanning areas in cybernetics, control theory, robotics, and neuroscience, and has led numerous multi-investigator multi-disciplinary research efforts with funding by the National Science Foundation and the Offi ce of Naval Research.