Scientific Computing and Research Consulting Services projects

Adaptive Resolution Imaging Sonar: Initial Exploration at the Shiawassee National Wildlife Refuge

The ARC Scientific Computing and Research Consulting Services supported a team of masters students from the School of Environment and Sustainability (SEAS) in developing a custom solution to conduct their research. 

The ARC Consulting Services team  provided support in grant writing, data pipeline development, machine learning training, and evaluation of a custom solution to identify fish in a sonar camera stream. Jetstream was used to store and process specialized data. The Consulting team also built on the ResNet-50 neural network architecture to train and run a neural network to locate and determine directionality of fish in the camera feed. Learn more about this SEAS and ARC collaboration. 

DDPG Training: Optimal Parameter Design in Circuits

ARC worked with a PhD student team to design a Deterministic Deep Policy Gradient agent to determine the optimal parameters for circuits to achieve the highest power efficiency. The DDPG training environment was a simulation block of a circuit model, which calculates the power efficiency given circuit parameters. This policy was implemented in MATLAB, using a training reward as a function of power efficiency.

Automating the Pipeline from Glomerular Detection to Morphometric Measures

ARC is currently collaborating with a research team from CSCAR, along with the U-M Medical School, to automate an approach to measure glomerular morphometry using neural networks along with segmentation and masking. This will allow for assessment of a large number of glomerular and other features from nephrectomy sections and kidney biopsies. We are using a convolutional neural network for deep learning classification of glomerular structures as well as segmentation for boundary detection and masking of glomerular areas. Learn more about the Medical School, CSCAR, and ARC collaboration. 

Alternative models for measuring biomass in practice

ARC consultants worked with a Master’s degree student team from the School for Environment and Sustainability (SEAS) comparing five modeling techniques — ordinary least squares (OLS) regression, linear fixed effects (FE), power law, random forest (RF), and support vector regression (SVR) — for estimating biomass from LiDAR data, which in turn can be used for predicting aboveground carbon storage using a limited amount of field data at the plot level to inform policy and preservation policy and practice. The comparison/assessment of the models was part of a project titled “Assessing and Communicating Climate and Water Ecosystem Services of the City of Ann Arbor Greenbelt Program.” The student team did not have funding, so ARC consultants helped it get a free allocation from the Open Science Grid (OSG) to run the computations and a free storage allocation from the OSiRIS project for ‘cloud’ storage accessible from both U-M and from OSG.

CV Parsing Project

ARC consultants worked with a research group in Ross to design a parsing algorithm in Python to extract information about education history, employment history, publications, and tenure status from input CVs. We implemented named entity recognition to find journals, publications, and universities referenced in a separate list. We also attempted to identify sections of the resume and place the identified publications and universities in the correct category (publications or employment/education). Overall, we were able to achieve moderate success with identifying publications and tenure status, and we made good progress on identifying education and employment history.

Natural Language Processing with Religious Sermons

Jacqueline Mattis, a professor at Rutgers and an associated faculty member in the School of Social Work, reached out to ARC Consulting to help write a letter of intent for a natural language processing project, hoping to identify demographic and ideological information from sermons given at various houses of worship in different geographical areas. For example, Mattis was interested in studying sermon content and model topics about racial injustice, and then relate that information to known demographic information of the house of worship in which the sermon was given. ARC Consulting worked to help her with a feasibility study and writing the letter of intent for the grant. We experimented with topic modeling with one such survey, including using an off-the-shelf model to convert speech to text and then an LDA model to model topics. We also tried using a semantic model to look at sentiments expressed in the sermon. This project was coded in Python using various scikit-learn packages.