July 10, 2015
The overall goal for the IR&D Program is to help position the Laboratory to make critical contributions to critical challenges in current and planned business areas through research and development. The Hart Prizes were established to recognize significant contributions to the Laboratory’s IR&D Program. This colloquium features the an IR&D Project recently selected in 2015 to receive the Hart Prize for Excellence in Research for FY2014.
Hart Prize for Research: MICrONS: Machine Intelligence from Cortical Networks
Investigators: Dean M. Kleissas, William Gray Roncal, Mark A. Chevillet, R. Jacob Vogelstein
The human brain provides an existence proof for high-performing, computationally-efficient intelligent algorithms. By understanding the structure and function of the brain, we may be able to reverse engineer its circuits, leading to greater understanding of neuropathologies and next-generation algorithms.
The challenge of high-resolution connectomics is to transform massive three-dimensional electron microscopy (EM) images of brain tissue into a neuronal connectivity map. Recent advances have enabled the complete imaging of individual neurons (i.e., graph vertices) and synapses (i.e., graph edges) over large regions. This capability allows us to directly estimate brain graphs at an unprecedented level of detail, and addresses a major knowledge gap in modern neuroscience. Image volumes of brain tissue are soon expected to reach the size of the cortical column proposed by Vernon Mountcastle (1 cubic millimeter), and will require new capabilities to analyze and store petabytes of raw data. To this end, our team recently built the first end-to-end fully automated pipeline to generate and assess brain graphs for EM data volumes, and are excited to share these results and future plans. This work has successfully transitioned from a multi-year REDD IRAD project to NIH and IARPA funded programs at the Lab.
William Gray Roncal is a Project Manager in the Research and Exploratory Development Department at The Johns Hopkins University Applied Physics Laboratory (APL). He is also a sixth year Ph.D. student in Computer Science at The Johns Hopkins University in Baltimore, Maryland.
Will applies computer vision algorithms to solve big data challenges. He has previous experience in diverse environments ranging from underwater to space, and currently works in connectomics, an emerging discipline within neuroscience that seeks to create a high-resolution map of the brain.
In 2005, Will received a Master of Electrical Engineering from the University of Southern California. He earned his Bachelor of Electrical Engineering Degree from Vanderbilt University in 2003. Will is a member of the Society for Neuroscience, Eta Kappa Nu, and Tau Beta Pi.
Will co-founded and leads the College Prep Program at APL, a free, all volunteer summer program to support and encourage underserved students who have the desire and academic potential to excel in college, but who lack the mentoring and resources necessary to succeed.
Dean Kleissas is a senior staff member in the Research and Exploratory Development Department at the Johns Hopkins University Applied Physics Laboratory. He received a B.S. in mechanical engineering and a B.S. in electrical and computer engineering from the University of Rochester in 2007. He also received a Master of Mechanical Engineering degree with a focus in robotics and control from The Johns Hopkins University in 2012.
Dean’s research interests span the multi-disciplined space of intelligent systems where he contributes to a variety of technical areas such as testing autonomous systems in complex environments, next-genearation retinal prosthetics, applications of machine learning and natural language processing to electronic medical records, and connectomics.