SIERRA (reSearch pIpEline for Reproducibility, Reusability, and Automation)

_images/architecture.png

SIERRA architecture, organized by pipeline stage. Stages are listed left to right, and an approximate joint architectural/functional stack is top to bottom for each stage. “…” indicates areas where SIERRA is designed via plugins to be easily extensible. “Host machine” indicates the machine SIERRA was invoked on.

SIERRA is a command line tool and plugin framework for:

  • Automating scientific research, providing faculties for seamless experiment generation, execution, and results processing.

  • Accelerating research cycles by allowing researchers to focus on the “science” aspects: developing new things and designing experiments to test them.

  • Improving the reproducibility of scientific research, particular in AI.

It supports a wide range of platforms, execution environments, and experiment input/output formats–see SIERRA Support Matrix for details. A short overview of the research automation SIERRA provides is here: SIERRA Pipeline. To get started, see Getting Started With SIERRA.

Citing SIERRA

If you use SIERRA and find it helpful, please cite the following paper:

@inproceedings{Harwell2022a-SIERRA,
 author = {Harwell, John and Lowmanstone, London and Gini, Maria},
 title = {SIERRA: A Modular Framework for Research Automation},
 year = {2022},
 isbn = {9781450392136},
 publisher = {International Foundation for Autonomous Agents and Multiagent Systems},
 address = {Richland, SC},
 booktitle = {Proceedings of the 21st International Conference on Autonomous Agents and Multiagent Systems},
 pages = {1905–1907},
 numpages = {3},
 keywords = {simulation, real robots, research automation, scientific method},
 location = {Virtual Event, New Zealand},
 series = {AAMAS '22}
 }

You can also cite the following DOI for the specific version of SIERRA used, to help facilitate reproducibility:

doi