<-Scope Go to ToC References->
The terms used in this standard whose first letter is capital have the meaning defined in Table 1.
Table 1 – Table of terms and definitions
| Term | Definition |
| Active Traceability Method | A Traceability Method that alters the NN Weights. |
| Algorithmic Integrity | The equivalence of the Traceability Data extracted from a modified NN and those extracted from an unmodified NN. |
| Computational Cost | The cost of injecting, Detecting, Decoding or Matching Traceability Data. |
| Detection | The process of finding the presence of a known watermark in a NN. |
| Decoding | The process of extracting the Payload from a watermarked NN. |
| Extraction | The process of computing the fingerprint from an NN. |
| Imperceptibility | A difference in the performance of an NN before and after the watermark embedding process. |
| Matching | The process of finding a fingerprint in a database that correspond to the fingerprint computed from an NN. |
| Means | Procedure, tools, dataset or dataset characteristics used to evaluate one or more of Computational Cost, Imperceptibility, or Robustness of a NN Traceability method. |
| Modification | The result of an attack performed during NN Traceability testing. |
| Neural Network | or Artificial Neural Network, a set of interconnected data processing nodes whose connections are affected by Weights. |
| NN Fingerprinting Method | A NN Passive Traceability method that extracts NN identification data from the NN Weights and matches it to a known repository. |
| NN Traceability | The possibility to identify the source and/or a potential Modification of a NN. |
| NN Watermarking Method | A NN Active Traceability method that injects Traceability Data into the Weights or the activation function of a NN to subsequently enable a Decoder/Detector to decode/detect the injected Traceability Data. |
| Parameter | A set of values characterizing Type and Intensity of a Modification, as used in Table 3. |
| Passive Traceability Method | A Traceability Method that does not alter the NN Weights. |
| Payload | The Symbols carried by a watermark. |
| Robustness | The ability of a NN Traceability method to withstand a Modification in terms of Detection, Decoding or Matching capability. |
| Secret Key | The data that the Traceability method requires to be kept secret. |
| Symbol | A binary, numerical, or string element in a Payload. |
| Tester | The user who evaluates a NN Traceability Method according to this Technical Specification. |
| Traceability | The possibility to trace the origin of data. |
| Traceability Data | The data to be inserted by the Active Traceability method or the result of the application of a Detection algorithm to an NN for a Passive Traceability Method. |
| Weight | The value used to multiply the connection between two nodes of a NN. |
The Terms used in this standard whose first letter is capital and are not already included in Table 1 are defined in Table 2.
| Term | Definition |
| Access | Static or slowly changing data that are required by an application such as domain knowledge data, data models, etc. |
| AI Framework (AIF) | The environment where AIWs are executed. |
| AI Module (AIM) | A data processing element receiving AIM-specific Inputs and producing AIM-specific Outputs according to according to its Function. An AIM may be an aggregation of AIMs. |
| AI Workflow (AIW) | A structured aggregation of AIMs implementing a Use Case receiving AIW-specific inputs and producing AIW-specific outputs according to the AIW Function. |
| Application Standard | An MPAI Standard designed to enable a particular application domain. |
| Channel | A connection between an output port of an AIM and an input port of an AIM. The term “connection” is also used as synonymous. |
| Communication | The infrastructure that implements message passing between AIMs |
| Component | One of the 7 AIF elements: Access, Communication, Controller, Internal Storage, Global Storage, Store, and User Agent |
| Conformance | The attribute of an Implementation of being a correct technical Implementation of a Technical Specification. |
| Conformance Tester | An entity Testing the Conformance of an Implementation. |
| Conformance Testing | The normative document specifying the Means to Test the Conformance of an Implementation. |
| Conformance Testing Means | Procedures, tools, data sets and/or data set characteristics to Test the Conformance of an Implementation. |
| Connection | A channel connecting an output port of an AIM and an input port of an AIM. |
| Controller | A Component that manages and controls the AIMs in the AIF, so that they execute in the correct order and at the time when they are needed |
| Data Format | The standard digital representation of data. |
| Data Semantics | The meaning of data. |
| Ecosystem | The ensemble of actors making it possible for a User to execute an application composed of an AIF, one or more AIWs, each with one or more AIMs potentially sourced from independent implementers. |
| Explainability | The ability to trace the output of an Implementation back to the inputs that have produced it. |
| Fairness | The attribute of an Implementation whose extent of applicability can be assessed by making the training set and/or network open to testing for bias and unanticipated results. |
| Function | The operations effected by an AIW or an AIM on input data. |
| Global Storage | A Component to store data shared by AIMs. |
| Internal Storage | A Component to store data of the individual AIMs. |
| Identifier | A name that uniquely identifies an Implementation. |
| Implementation | 1. An embodiment of the MPAI-AIF Technical Specification, or
2. An AIW or AIM of a particular Level (1-2-3) conforming with a Use Case of an MPAI Application Standard. |
| Implementer | A legal entity implementing MPAI Technical Specifications. |
| ImplementerID (IID) | A unique name assigned by the ImplementerID Registration Authority to an Implementer. |
| ImplementerID Registration Authority (IIDRA) | The entity appointed by MPAI to assign ImplementerID’s to Implementers. |
| Interoperability | The ability to functionally replace an AIM with another AIW having the same Interoperability Level |
| Interoperability Level | The attribute of an AIW and its AIMs to be executable in an AIF Implementation and to:
1. Be proprietary (Level 1) |
| Knowledge Base | Structured and/or unstructured information made accessible to AIMs via MPAI-specified interfaces |
| Message | A sequence of Records transported by Communication through Channels. |
| Normativity | The set of attributes of a technology or a set of technologies specified by the applicable parts of an MPAI standard. |
| Performance | The attribute of an Implementation of being Reliable, Robust, Fair and Replicable. |
| Performance Assessment | The normative document specifying the Means to Assess the Grade of Performance of an Implementation. |
| Performance Assessment Means | Procedures, tools, data sets and/or data set characteristics to Assess the Performance of an Implementation. |
| Performance Assessor | An entity Assessing the Performance of an Implementation. |
| Profile | A particular subset of the technologies used in MPAI-AIF or an AIW of an Application Standard and, where applicable, the classes, other subsets, options and parameters relevant to that subset. |
| Record | A data structure with a specified structure |
| Reference Model | The AIMs and theirs Connections in an AIW. |
| Reference Software | A technically correct software implementation of a Technical Specification containing source code, or source and compiled code. |
| Reliability | The attribute of an Implementation that performs as specified by the Application Standard, profile and version the Implementation refers to, e.g., within the application scope, stated limitations, and for the period of time specified by the Implementer. |
| Replicability | The attribute of an Implementation whose Performance, as Assessed by a Performance Assessor, can be replicated, within an agreed level, by another Performance Assessor. |
| Robustness | The attribute of an Implementation that copes with data outside of the stated application scope with an estimated degree of confidence. |
| Scope | The domain of applicability of an MPAI Application Standard |
| Service Provider | An entrepreneur who offers an Implementation as a service (e.g., a recommendation service) to Users. |
| Standard | The ensemble of Technical Specification, Reference Software, Conformance Testing and Performance Assessment of an MPAI application Standard. |
| Technical Specification | (Framework) the normative specification of the AIF. (Application) the normative specification of the set of AIWs belonging to an application domain along with the AIMs required to Implement the AIWs that includes: 1. The formats of the Input/Output data of the AIWs implementing the AIWs. 2. The Connections of the AIMs of the AIW. 3. The formats of the Input/Output data of the AIMs belonging to the AIW. |
| Testing Laboratory | A laboratory accredited to Assess the Grade of Performance of Implementations. |
| Time Base | The protocol specifying how Components can access timing information |
| Topology | The set of AIM Connections of an AIW. |
| Use Case | A particular instance of the Application domain target of an Application Standard. |
| User | A user of an Implementation. |
| User Agent | The Component interfacing the user with an AIF through the Controller. |
| Version | A revision or extension of a Standard or of one of its elements. |
| Zero Trust | A model of cybersecurity primarily focused on data and service protection that assumes no implicit trust. |
