The IIC is reviewing resumes for summer internships. Please submit
your resumes with a cover letter stating your area of research.
Interested students may check the URL below.
http://iic.harvard.edu/employment/index.html
2008 Summer Internships
Astronomical Medicine
The Initiative in Innovative Computing at Harvard seeks a qualified
intern to implement software algorithms for segmenting and searching
for features in cross-discipline scientific imaging. The IIC has
several projects that involve classifying, retrieving, and
visualizing regions of images that meet particular statistical
criteria. These projects include a medical imaging/astronomy
collaboration, a time series analysis center, and an effort to
segment large amounts of neural tissue from electron microscopy data.
This intern would focus intensively on one application domain, the
use of the open-source Insight Toolkit (ITK) for segmentation of
astronomy data (potentially including techniques such as statistical
classification, interactive editing, and visual display of results),
while providing guidance to the rest of IIC on the development of
cross-discipline tools and techniques for scientific image
segmentation. Image data for this project comes from multiple data
channels acquired at different wavelengths, using different
instruments, and will be segmented according to intensity-, region-,
and shape-based criteria.
A successful applicant should have working knowledge of image
segmentation algorithms and software frameworks at the Master's
Degree level or beyond. Working knowledge of ITK and C++ a strong
plus. Experience with developing interactive software tools,
visualization applications, or search algorithms also a plus.
Design Group
The intern will be involved with the development of Drupal-based web
sites which will require the student to be familiar with HTML, CSS,
PHP. Familiarity with CVS/SVN, Drupal and MySQL and Photoshop is
desirable but not necessary.
At the end of the summer, it will be possible to show progress on a
number of different web sites both internal and external. New
features and capabilities will be supported, most notably on the IIC
website to support internal documents and workflow.
The student will learn best practices in site development and
deployment and get experience in HTML, CSS, PHP and Drupal.
Connectome
The goal of the project is to do research in automatic extraction of
axons from electron microscopy images of brain tissue. We will
investigate using shape-based surface evolution models for the
segmentation of EM volume data of neural tissue. We will develop
active contour models using ITK/VTK that searches for tubular
structures using Higher Order Active Contours. The model minimizes an
energy functional that favors branching cylindrical shapes over others.
The intern will develop a new method for automatic extraction of
axons in EM volume data. The intern will implement his ideas and
algorithms in software. All findings will be properly documented in a
final internal report. Tracking axons in EM volume data is one of the
key components of the Connectome project and will have substantial
impact in neuroscience.
Scientist Collaboration Framework
Computer science graduate student with experience in web technologies
including PHP, AJAX & HTML. Familiarity with emerging Semantic Web
technologies such RDF, OWL and triple-stores. Skilled at programming
and software development methodologies.
Create a semantic module for bibliographies in Drupal (content
management system). Such a system would allow easy import/export of
bibliographies from external sources (such as EndNote, RefMan or
Bibtex) as well as allow the user to easily create a new entry. The
framework will allow tagging of the bibliography with keywords from
controlled vocabularies. Users of the framework will be able to
perform high precision knowledge mining on these bibliographies.
Drupal is increasingly being adopted as the content management of
choice by scientific communities (e.g. IIC, HSCI). Bibliography
maintenance, annotation and mining is one of the challenges for such
a web-based community and the goal of this module is to address that
challenge.
Time Series Center
The IIC Time Series Center is collaborating with scientist, computer
scientists, and statisticians to develop a-priori algorithms for the
discovery of anomalies, similarities and dis-similarities, and
patterns in time series. Our research combines work in computer
science, astronomy, numerical and symbolic algorithms, web services,
and databases. Early classification and early detection of anomalous
events in time series.
Machine learning techniques such as Support Vector Machine (SVM) and
k Nearest Neighbors (kNN) can be applied to classify time series by
applying a distance metric. Such techniques have been applied
successfully in different fields, e.g. in astronomy to classify
periodic stars. In this project we want to explore methods on
streaming data for early warning systems. We will apply existing
probabilistic classification methods as well as develop new
approaches in order to identify automatically anomalous events.
Methods that produce high accuracy will be applied to existing data
at the Time Series Center.
We are looking for a student with a strong background in Computer
Science and strong understanding of machine learning and
understanding of time series analysis.
_______________________________________________
iic-seminars mailing list
iic-seminars(a)calists.harvard.edu
http://calists.harvard.edu/mailman/listinfo/iic-seminars