<div dir="ltr"><div class="gmail_quote"><div dir="ltr"><div><div><div><div class="gmail_quote"><span></span><div dir="ltr"><div><span>I am pleased to announce the release of the Machine Intelligence from Cortical Networks (MICrONS) program Broad Agency Announcement (BAA). <span style="color:black">MICrONS
seeks to revolutionize machine learning by reverse-engineering the algorithms
of the brain. <span> </span>The program is expressly
designed as a dialogue between data science and neuroscience, in which participants will have the
unique opportunity to pose biological questions with the greatest potential to
advance theories of neural computation and obtain answers through carefully
planned experimentation and data analysis.<span>
</span>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.<span> </span>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.<span></span></span> <br><br></span>The program overview is copied below; the full text of the announcement (and any future versions) is available from the BAA link at <a href="http://www.iarpa.gov/index.php/research-programs/microns/microns-baa" target="_blank">http://www.iarpa.gov/index.php/research-programs/microns/microns-baa</a>. For your convenience, a PDF of the announcement is also attached to this email. All questions about the program and/or BAA must be submitted to <a href="mailto:dni-iarpa-baa-14-06@iarpa.gov" target="_blank">dni-iarpa-baa-14-06@iarpa.gov</a> by February 9, 2015. Full proposals must be submitted through the <a href="https://iarpa-ideas.gov" target="_blank">IARPA IDEAS</a> system by March 13, 2015. <i>Do not </i>send any questions or proposal submissions to me directly. <br><span><br></span><span>Please disseminate this information widely. </span><span><span>Offerors need not be U.S. citizens or residents to apply or receive funding. </span></span><br><span><br>Thank you,<br>Jacob<br><br>R. Jacob Vogelstein, Ph.D.<br>Program Manager<br>ODNI/IARPA<div><a href="http://go.usa.gov/FPPJ" target="_blank">http://go.usa.gov/FPPJ</a></div><br>--------------------------------------------------------<span style="color:black"><br><span></span></span></span></div><span><p></p><p><span style="color:black"><span></span></span></p>
<div>
<h3><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Toc408388859"><span style="color:black">Introduction</span></a><span style="color:black"></span></h3>
<p><span style="color:black">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.<span> </span>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.<span> </span>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.<span>
</span>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. </span></p>
<h3><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Ref275032838"></a><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Toc408388860"><span><span style="color:black"><span><span style="font:7pt "Times New Roman""> </span></span></span><span style="color:black">Background</span></span></a><span><span style="color:black"> </span></span><span style="color:black"></span></h3>
<p><span style="color:black">Many
contemporary theories of cortical computing suggest that the brain performs
common sensory information processing tasks—such as detection and recognition of
visual objects, sounds, and odors—with algorithms that progressively transform
data through a series of operations, or “stages.” <span> </span>Each stage of processing is further theorized
to occur within a discrete region of cortex.<span>
</span>Although different theories suggest different mathematical bases for
computation, it is commonly believed that neural algorithms employ data
representations, transformations, and learning rules that are conserved across
stages.<a href="#14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__ftn1" name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__ftnref1" title=""><span><span><span><span style="font-size:12pt;font-family:"Times New Roman","serif";color:black"></span></span></span></span></a> <span></span>It should therefore be possible to apprehend
the neural computations underlying information processing (at least within a
given sensory modality<a href="#14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__ftn2" name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__ftnref2" title=""><span><span><span><span style="font-size:12pt;font-family:"Times New Roman","serif";color:black"></span></span></span></span></a>) by
interrogating a small fraction of the entire cortex, so long as that fraction is
judiciously selected to contain sufficient evidence of the representations,
transformations, and learning rules of the algorithm(s) to which it
contributes.</span></p>
<p><span style="color:black">Neuroscience
has a long history of inspiring innovation in machine learning, starting with
the seminal work of McCulloch and Pitts in 1943.<span> </span>This influence is evident even in today’s
state of the art “deep learning” systems, which are loosely modeled on
hierarchical visual processing systems in the primate brain.<span> </span>However, the rate of effective knowledge
transfer between neuroscience and machine learning has been slow because of
divergent scientific priorities, funding sources, knowledge repositories, and
lexicons.<span> </span>As a result, very few of the
ideas about neural computing that have emerged over the past few decades have
been incorporated into modern machine learning algorithms.<span> </span></span></p>
<p><span style="color:black">Previous
attempts to foster collaboration between neuroscience and machine learning have
been stymied in part by gaps in our knowledge about the brain.<span> </span>The majority of what is known about the brain
today regards its operation at the “micro” scale (one or a few neurons) and the
“macro” scale (hundreds of thousands or millions of neurons), and some of this
information is indeed reflected in the design of leading artificial neural
networks.<span> </span>In contrast, much less is
known about the “mesoscale” cortical circuits (hundreds to tens of thousands of
neurons) that implement the specific data representations, transformations, and
learning rules of cortical information processing algorithms, and these are
therefore absent from (or speculative in) existing machine learning
solutions.<span> </span>It is likely that explicit
knowledge and use of these computations is required to move beyond the current
generation of “neurally-inspired” machine learning algorithms.<span> </span></span></p>
<h3><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Toc408388861"></a><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Ref400439142"><span><span style="color:black">Program Synopsis</span></span></a><span style="color:black"></span></h3>
<p><span style="color:black">The MICrONS
program aims to create novel machine learning algorithms that use neurally-inspired
architectures <i>and</i> mathematical
abstractions of the representations, transformations, and learning rules
employed by the brain to achieve brain-like performance.<span> </span>To guide the construction of these
algorithms, performers will conduct targeted neuroscience experiments that
interrogate the operation of mesoscale cortical computing circuits, taking
advantage of emerging tools for high-resolution structural and functional brain
mapping.<span> </span>The program is designed to
facilitate iterative refinement of algorithms based on a combination of practical,
theoretical, and experimental outcomes: performers will use their experiences
with the algorithms’ design and performance to reveal gaps in their
understanding of cortical computation, and will collect specific neuroscience
data to inform new algorithmic implementations that address these limitations. <span> </span>Ultimately, as performers incorporate these
insights into successive versions of the machine learning algorithms, they will
devise solutions that can perform complex information processing tasks aiming
towards human-like proficiency.<span> </span></span></p>
<h3><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Toc408388862"></a><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Ref406506171"></a><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Ref279612573"><span><span><span style="color:black">Program Structure</span></span></span></a><span style="color:black"> </span></h3>
<p><span style="color:black">MICrONS is
organized in three phases, totaling five years in
duration.<span> </span>During each phase, performers conduct
targeted neuroanatomical and neurophysiological studies to inform their
understanding of the cortical computations underlying sensory information
processing and, concurrently, create neurally-derived machine learning
algorithms that perform similar functions. <span> </span>Performers motivate their experimental and
algorithmic designs by formulating and updating a conceptual model or
“theoretical framework” for neural information processing in a given sensory
modality.<span> </span>They use computational neural
models (i.e., </span>executable mathematic or algorithmic simulations of
neurons and neural circuits<span style="color:black">) to establish
a correspondence between the computations performed by biological wetware and
the computations employed by their machine learning software.<span> </span>Each phase ends with an information
processing challenge that assesses how well the new algorithms perform on
increasingly challenging machine learning tasks: similarity discrimination in
Phase 1, generalization and classification in Phase 2, and invariant
recognition in Phase 3.<span> </span>Performers use
the results of their experiments in each phase to guide their development of
improved algorithms in the subsequent phase (in Phase 1, performers base their algorithms
on the existing neuroscience literature). </span><span style="color:black"> </span></p>
<h3><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Toc408388863"></a><a name="14ae94551cb4b667_14ae942f2d046d85_14acb3be21100ccb__Ref279849771"><span><span style="color:black">Technical Areas</span></span></a><span style="color:black"> </span></h3>
<p class="MsoNormal"><span style="color:black">The MICrONS
program comprises three Technical Areas (TAs). <span> </span>Although IARPA anticipates receiving a number
of holistic proposals responding to all three TAs, it recognizes that some
prospective offerors may have capabilities in only a subset of the overall
program scope, and wishes to maximize its opportunity to leverage these
capabilities.<span> </span>Therefore, offerors may
choose to propose to one, two, or all three TAs. <span> </span>Because achieving MICrONS program goals will
require significant collaboration across all three TAs, offerors who propose to
only one or two TAs should be prepared to work closely with performers in the remaining
TAs.<span> </span>The TAs in MICrONS are defined as
follows: <br></span></p><p class="MsoNormal"><span style="color:black"><br></span></p>
<p style="margin-top:3pt"><span style="font-family:Symbol;color:black"><span>·<span style="font:7pt "Times New Roman"">
</span></span></span><span style="color:black">TA1
– experimental design, theoretical neuroscience, computational neural modeling,
machine learning, neurophysiological data collection, and data analysis;</span></p>
<p style="margin-top:3pt"><span style="font-family:Symbol;color:black"><span>·<span style="font:7pt "Times New Roman"">
</span></span></span><span style="color:black">TA2
– neuroanatomical data collection; and</span></p>
<p style="margin-top:3pt"><span style="font-family:Symbol;color:black"><span>·<span style="font:7pt "Times New Roman"">
</span></span></span><span style="color:black">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.<span> </span></span></p>
<p class="MsoNormal"><span style="color:black">Success in
the MICrONS program will require extensive communication and cooperation between
performers in all three TAs within or across teams.<span> </span>For example, in TA2, performers must collect
neuroanatomical data about the same brain regions <i>in the same brain specimens</i> that are used in TA1 for
neurophysiological studies; in TA3, performers must reconstruct neural circuits
from the data collected in TA2; and in TA1, performers must analyze the neural
circuits generated in TA3 and use the resulting insights in formulating their
machine learning algorithms and theoretical frameworks.<span> </span>All offerors are therefore required to
include in their proposal a detailed management plan </span><span style="color:black"> and a detailed description of how
their proposed technical approach in one or more TAs is likely to impact the
other TAs</span><span style="color:black">.<span>
</span></span></p>
</div>
<br><br></span></div>
</div><br></div></div></div></div>
</div><br></div>