US intelligence officials have set in motion a five-year project to spark progress in machine learning by reverse-engineering the algorithms of the human brain. The Intelligence Advanced Research Projects Agency (IARPA) recently put out a call for innovative solutions with the greatest potential to advance theories of neural computation as part of the Machine Intelligence from Cortical Networks (MICrONS) program. The agency, known for its funding of high-risk/high-payoff research in support of national intelligence, is ultimately looking to facilitate the development of synthetic systems with brain-like performance and proficiency.

In a just-issued broad agency announcement, IARPA lays out its strategy for fostering multidisciplinary approaches at the intersection of data science and neuroscience that increase scientific understanding of the cortical computations underlying neural information processing. Although there has been much progress in modeling machine learning algorithms after neural processes, the brain remains far better-suited for a host of detection and recognition tasks.

The agency sees the emerging research area of neurally-inspired machine learning as crucial for closing the performance gap between software and wetware.

“Despite significant progress in machine learning over the past few years, today’s state of the art algorithms are brittle and do not generalize well,” the proposal’s authors contend. “In contrast, the brain is able to robustly separate and categorize signals in the presence of significant noise and non-linear transformations, and can extrapolate from single examples to entire classes of stimuli. This performance gap between software and wetware persists despite some correspondence between the architecture of the leading machine learning algorithms and their biological counterparts in the brain, presumably because the two still differ significantly in the details of operation. The MICrONS program is predicated on the notion that it will be possible to achieve major breakthroughs in machine learning if we can construct synthetic systems that not only resemble the high-level blueprints of the brain, but also employ lower-level computing modules derived from the specific computations performed by cortical circuits.”

The MICrONS program consists of three Technical Areas (TAs), defined as follows:

TA1 – experimental design, theoretical neuroscience, computational neural modeling, machine learning, neurophysiological data collection, and data analysis;

TA2 – neuroanatomical data collection; and

TA3 – reconstruction of cortical circuits from neuroanatomical data and development of information technology systems to store, align, and access neural circuit reconstructions with the associated neurophysiological and neuroanatomical data.

“Over the course of the program, participants will use their improving understanding of the representations, transformations, and learning rules employed by the brain to create ever more capable neurally-derived machine learning algorithms,” the IARPA proposal further explains. “Ultimate computational goals for MICrONS include the ability to perform complex information processing tasks such as one-shot learning, unsupervised clustering, and scene parsing, aiming towards human-like proficiency.”

MICrONS is set to run from September 2015 through September 2020. A summary of the scientific and technical objectives of each program phase as well as a very comprehensive set of metrics for each of the technical areas are detailed in the full 70-page solicitation [PDF].