NIST logo
*

Data Science Program

Summary:

The NIST Information Technology Laboratory (ITL) has over 25 years of experience measuring and evaluating technology that is used to process, analyze, and derive knowledge from various structured and unstructured data, including text, audio/voice, imagery, video, and multimedia. ITL and the Information Access Division (IAD) within ITL have extended our current efforts by forming a new, cross-cutting data science program, focused on driving advancements in big data analytics. This focus, alongside our collaboration with industry, academia, and government, will accelerate development of these technologies as well as enable faster transition into government applications and the commercial market place.

As part of this program, ITL and IAD will create and manage a big data analytics evaluation series and develop a model of data science through which we can measure various attributes, e.g., component classes, component-level performance, end-to-end performance, flow analyses, and propagation of uncertainty. This approach will define metrology supporting objective comparison of approaches, identify key areas for improvement in big data analytics, and focus research community critical mass on the hard problems.

Our first data science symposium is being planned for March 2014, which will be set apart from other related symposia by our emphasis on advancing the technology through community-wide evaluations, metrics, reference datasets, and challenge tasks.

Description:

Current Data Science measurement and evaluation activities:

Modeling Data Science: The complexities inherent to big data analytics require new methods and models to inform novel evaluation paradigms and measurement methods. NIST is collaborating with data science researchers and practitioners to develop models that will support the metrology necessary to drive data science forward.

Data Analytics Evaluation Series: In 2013, the NIST Information Technology Laboratory and Information Access Division began investigation into requirements for a new big data analytics evaluation series.

Data Science Symposium Series: The inaugural NIST Data Science Symposium is being planned for March 4-5 2014

Lead Organizational Unit:

itl
Contact

NIST maintains a general mailing list for our Data Science Measurement and Evaluation program. To join this list, please email us using mailto:datascience-list-request@nist.gov?subject=subscribe

Relevant information is posted to this list. If you have any question for NIST related to our data science program, please email us at

datascience@nist.gov.