Winner of the 2011-2012 inaugural Neukom Institute/IQBS CompX Faculty Grants Program for Dartmouth faculty has been announced with an award of up to $20,000 for a one-year project.
Co-sponsored by the Neukom Institute and the Institute for Quantitative Biomedical Sciences(Geisel), the program is focused on funding computational biomedical research in bioengineering, bioinformatics, biostatistics, biophysics or other related areas across the campus and professional schools.
Dartmouth College faculty including the undergraduate, graduate, and professional schools were eligible to apply for this competitive grant. The inaugural winner is:
There are 30 million people in the United States who suffer from sensorineural hearing loss. For those with the most severe form of this condition, auditory communication can be restored only with a neuroprosthetic aid called a cochlear implant. This is a nanoelectronic device that directly stimulates the eighth cranial nerve with encoded pulses of electricity, which are then transported to the auditory center, ultimately evoking a sensation of hearing for the user.
Three decades after the first device was implanted, these “bionic ears” are by far the most successful neuroprosthetic to date. However, their capabilities are severely overwhelmed when faced with complex auditory environments, where a sound of interest (e.g. a person’s voice) is heard in competition with other background noises. Such everyday scenarios like a busy restaurant or a lively conversation will unfortunately confound a cochlear implant and render it effectively useless.
Today, digital signal processing algorithms now exist that can separate an individual auditory stream from other noises, provided the algorithms are continually and rapidly fed the appropriate signaling cues. The most effective way for such algorithms to assess the signaling cues is by directly coupling to the user’s brain, via electroencephalography (EEG). To interpret and analyze these brain signals in conjunction with the auditory scene in question, a conventionally implemented algorithm requires large amounts of processing power and must be run on a workstation computer or on a high-performance laptop. It is completely unsuitable for an embedded application like a neuroprosthetic. This is why cochlear implants are to this day still limited in their capabilities.
The Analog Signal Processing Lab in Dartmouth’s Thayer School of Engineering is actively developing new processing paradigms that are orders of magnitude more power and area efficient than the conventional digital approaches. In collaboration with Dr. James Saunders, a neurotologist at the DHMC, our goal is to achieve auditory target-background noise separation on a system that has a 20 millimeter-squared footprint. This is about 1/10th the area of a US 10 cent coin. The system would consume so little energy that it could be powered by the mechanical vibrations that are generated by the user’s natural movements. Such a system would propel today’s cochlear implants to a revolutionary level of cyberphysical sophistication.
Hanson, Valerie, and Kofi Odame. Real-time source separation on a field programmable gate array platform. In Conference proceedings:... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference (Vol. 2012, pp. 2925-2928).
Winners of the 2011-2012 Neukom Institute CompX Faculty Grants Program for Dartmouth faculty have been announced with awards of up to $20,000 for one-year projects.
The program seeks to fund both the development of novel computational techniques as well as the application of computational methods to research across the campus and professional schools.
Dartmouth College faculty including the undergraduate, graduate, and professional schools were eligible to apply for these competitive grants. This years winners are:
Agent-based modeling (ABM) allows a researcher to observe a complex, dynamic system as it emerges from individual interactions of “agents” into a social network with particular characteristics. For example, a basic ABM program might be designed to initialize a number of “agents” at random locations in a virtual space andthen parameterize each agent to be attracted to other agents. After a number of iterations, the agents will naturally cluster themselves into several highly concentrated “cities”. Depending on the parameters used, an ABM researcher can model a wide range of systems, including human social behavior (Heath, Hill & Ciarallo 2009).
ABM is now being used in the social sciences, biology, medicine, ecology, economics, public policy, and many other fields of study (Heath et al. 2009). Despite the rapid growth of ABM in many academic disciplines over the last decade, we note that linguistics is underrepresented, especially in the subfields of dialect research and endangered language research. The relative dearth of ABM in linguistic research is surprising since a great deal of current linguistics literature visualizes human language as a complex, dynamic system that emerges from the agentive behavior of individual speakers in a community (Eckert 2005; Bybee 2007; Kretzschmar 2009).
Therefore, our proposal explores an expanded use of ABM in the field of linguistics, addressing two specific research questions: (1) Can ABM derive the classic principles of dialect research that were originally developed through observational sociolinguistics? (2) What can ABM tell us about endangered languages and how to revitalize the languages of threatened minority communities?
Dartmouth NOW: Dartmouth Linguists Remap Boundary Between East, West (VPR)
Neighborhood racial composition is often conceived as running along a continuum from “segregated” at one extreme to “diverse” at the other. In this proposal, we seek to employ advanced computational methods – geospatial technology and interactive web-based mapping – to visualize and appraise both segregation and diversity simultaneously.
The racial composition of American communities is both spatially heterogeneous and, in many places, rapidly changing. Much of this change is driven by the arrival of new immigrant populations. Does this process lead to an increase in neighborhood diversity – or to increased residential segregation? Collaborating with colleagues Mark Ellis (University of Washington) and Steven Holloway (University of Georgia), our recent work has led to the development of a new taxonomy of neighborhood racial composition (Holloway et al. 2011; Wright et al. 2011). Instead of positioning neighborhoods along a unidimensional scale, this new classification scheme incorporates both the degree of diversity within a neighborhood, and, for low- or moderate-diversity neighborhoods, the identity of the numerically dominant racial group. Thus, we could speak of communities as “black dominant and moderately diverse” or “Asian dominant and not diverse”.
We have constructed a prototype website that allows users to visualize spatial patterns of segregation and diversity, using census data from 1990 and 2000, for 50 US metropolitan areas (http://www.mixedmetro.com). We now propose to expand this resource – spatially, temporally, and technologically. In this project, we will add new data from the 2010 census, and expand the scope to include all census tracts in all 50 states and to feature segregation and diversity in metropolitan areas that have over 1 million people. We will also supplement this “web-based atlas” with an online geodatabase, allowing users to use our information products in their own work.
Dartmouth NOW: Racial Diversity Increases but Segregation Persists
Last Updated: 10/8/15