Appendix C: Glossary

Appendix C: Glossary

Application programming interface that defines the kinds of calls or requests that can be made, how to make them, the data formats that should be used, the conventions to follow, etc.

Autonomous vehicle:
A vehicle capable of sensing its environment and operating without human involvement.

Black box:
Any complex piece of equipment, typically a unit in an electronic system, with contents that are mysterious to the user.

A computer program designed to simulate conversation with human users, especially over the internet.

Civil Society Organizations (CSOs):
Non-state, not-for-profit, voluntary entities formed by people in the social sphere that are separate from the State and the market. CSOs represent a wide range of interests and ties. They can include community-based organizations as well as non-governmental organizations (NGOs).

To spread (something, especially information) widely.

A parameter whose value is used to control the learning process.

Impacted non-users:
Persons or organizations impacted by a system even though they did not directly use it (i.e., a banking customer [nonuser] rejected for a home loan by a 3rd party algorithm [ML System] owned by some organization external to the bank and possibly implemented without the knowledge of the banking customer).

Industry norm:
An authoritative standard.

Intellectual property:
Any work or invention that is the result of creativity, such as a manuscript or a design, to which one has rights and for which one may apply for a patent, copyright, trademark, etc.

Iterative process:
A series of steps that are repeated, tweaking and improving the product with each repetition.

Model card:
A type of documentation artifact accompanying a given model that details the model itself, its intended uses, potential limitations, training parameters, datasets used, experimental information, and model evaluation results. (Definition from Hugging Face)

Multistakeholder process:
A process which aims to bring together the primary stakeholders such as businesses, civil society, governments, research institutions and non-government organizations to cooperate and participate in the dialogue, decision-making and implementation of solutions to common problems or goals.

Natural Language Processing, or NLP for short, is broadly defined as the automatic manipulation of natural language, like speech and text, by software.

Objective and Key Result (OKR):
A collaborative goal-setting tool used by teams and individuals to set challenging, ambitious goals with measurable results. OKRs are how you track progress, create alignment, and encourage engagement around measurable goals

Reidentification risk:
The practice of matching anonymous data (also known as de-identified data) with publicly available information, or auxiliary data, in order to discover the individual to which the data belong to. This is a concern because companies with privacy policies, health care providers, and financial institutions may release the data they collect after the data has gone through the de-identification process.

The extent to which consistent results are obtained when an experiment is repeated.

Trade secrets:
Defined by the United States Patent and Trademark Office as information that has either actual or potential independent economic value by virtue of not being generally known, has value to others who cannot legitimately obtain the information, and is subject to reasonable efforts to maintain its secrecy.