Four Professors Win 2016 NSF CAREER Awards
Four Columbia Engineering assistant professors have won Faculty Early Career Development (CAREER) awards from the National Science Foundation: Allison Bishop (Computer Science), Matei Ciocarlie, (Mechanical Engineering), Pierre Gentine (Earth and Environmental Engineering), and Nima Mesgarani (Electrical Engineering). CAREER grants are the NSF’s most prestigious award in support of the early career-development activities of junior faculty who most effectively integrate research and education within the context of the mission of their organization.
Bishop, who is also a member of the Data Science Institute, will use the award to continue developing tools to share and protect sensitive information as threats to personal privacy and data security grow in the big data era. She is looking specifically at integrating recent advances in lattice cryptography with her advances in designing security reductions in order to provide strong arguments for the security of highly flexible and customizable cryptographic systems. Flexibility is a key challenge, and Bishop hopes to build secure cryptographic systems that can accommodate various levels of access to data, thus allowing different people to access different data within the same data source. She will also use the grant to provide an entry point and training ground for emerging young scientists of all ages, including elementary school children for whom she is writing a book that uses an interactive, fantasy format to introduce mathematical reasoning.
Ciocarlie focuses on developing versatile manipulation and mobility in robotics, in particular on building dexterity into robotic hands. He will use the CAREER award to explore the science of manipulation, resulting in better tools to assess the mechanisms and methods for planning and executing manipulation tasks. He hopes that by optimizing hand designs jointly with planning and control algorithms, using combined quality metrics, he will be able to build versatile mechanisms that are ready for use right after assembly, and already programmed for a wide range of useful tasks. His concept is a departure from the so-called “build, then program” paradigm, where the hardware and software components of a manipulator are developed either sequentially or independently. His group is working on a range of applications that are part of everyday life, from versatile automation in manufacturing and logistics to mobile manipulation in unstructured environments to assistive and rehabilitation robotics in health care.
Gentine studies land and atmosphere interactions and the inherent feedback between the two systems. He will use the award to look at better modeling turbulent fluxes and account for the largest and most efficient eddies in his team’s representation of turbulent exchange at the Earth's surface. Turbulence controls the exchange of heat, moisture, air, and passive chemical tracers such as CO2 flow between the land and the atmosphere. Accurate model representation of such turbulent fluxes is essential for precise hydrologic, weather, and climate predictions. Current model representation of turbulent fluxes assumes that most eddies’ transport can be explained by local observations and parameters in models. But variability in the horizontal (due to changes in the surface characteristics) and in the vertical (due to eddies that span an unusually large vertical extent) can invalidate these assumptions. In this project, he will test the latter large eddies effect using a combination of high-resolution turbulence models and observations run on supercomputers at Yellowstone. His goal is to better account for the largest—most efficient—eddies in the representation of turbulent exchange at the surface. He hopes this work will improve the way researchers measure and model surface fluxes and better predict evaporation, heating and carbon fluxes.
Mesgarani’s CAREER project will explore a new research direction for his group: computational modeling of neural networks. The recent parallel breakthroughs in deep neural network models and neuroimaging techniques have significantly advanced the current state of artificial and biological computing. However, there has been little interaction between these two disciplines, resulting in simplistic models of neural systems with limited prediction, learning, and generalization abilities. Mesgarani’s goal is to create a coherent theoretical and mathematical framework to understand the computational role of distinctive features of biological neural networks, their contribution to the formation of robust signal representations, and to model and integrate them into the current artificial neural networks. These new bio-inspired models and algorithms will have adaptive and cognitive abilities, will better predict experimental observations, and will advance the knowledge of how the brain processes speech. In addition, the performance of these models should approach human abilities in tasks mimicking cognitive functions, and will motivate new experiments that can further impose realistic constraints on the models.
—by Holly Evarts