Improving Media Analysis and Research by Developing a Model Training Pipeline

  • Post category:Projects

Media Cloud is an open-source platform for media analysis and research. Users can query the online news archive to investigate media attention and language. To expand the available toolchain, this team will build a model training pipeline that guides users in specifying an initial query, downloading and labeling sample content, and iteratively training a classifier model. 

Continue ReadingImproving Media Analysis and Research by Developing a Model Training Pipeline

Analyzing Drone-Captured Images of African Elephants

  • Post category:Projects

Save the Elephants is an organization that works to secure a future for elephants. This team is working with Save the Elephants to analyze drone-captured images of African elephants who are currently being negatively impacted by doubt and climate change. The team will use these images to develop models that help track elephants, monitor their health information, and prevent human-elephant conflict. 

Continue ReadingAnalyzing Drone-Captured Images of African Elephants

Identifying Patterns in Unity’s GPU Usage

  • Post category:Projects

The Unity Research Computing Platform accelerates research from laptop-size to 1,000’s of cores and 100’s of graphics processing units (GPUs), and it has about 1400 GPUs that are often underutilized by user jobs. This team will develop an automated workflow to identify patterns in GPU utilization, summarize common issues, and categorize user profiles, for the purpose of providing timely intervention and effective user outreach.   

Continue ReadingIdentifying Patterns in Unity’s GPU Usage

Generating Metrics for High-Performance Computing Clusters

  • Post category:Projects

Partner: Unity DS4CG 2023. Unity is a collaborative, multi-institutional high-performance computing cluster, primarily used for research computing. The Unity project focused on generating useful metrics and analysis for Unity by building a pipeline to a database that could power a live dashboard for Unity’s admin staff. Metrics included unnecessarily idle GPUs, daily and weekly node usage, total resource usage, and wait time. Additionally, a prediction model for wait time at job submission time was built.

Continue ReadingGenerating Metrics for High-Performance Computing Clusters