Exploring large datasets can not only be challenging but also time consuming. The data science workflow encompasses ETL - where data is fetched from disparate places, data preparation, machine learning model creation and training, and an evaluation of predicted results. While CPU based computng has historically excelled in exploring data lakes, compute time can often be slow - preventing a developer from having a “conversation” with her data.
During this webinar, we will introduce RAPIDS: a community driven open source set of software libraries that GPU accelerates the data science workflow. By focusing on ease of use, RAPIDS bridges the gap between high performance compute and typical software packages used by the data science community, like Pandas or Scikit-Learn. We will also discuss how some federal stakeholders are putting RAPIDS to work for cyber security applications.
