Linked.Archi ML-Enabled Systems Extension
language en

Linked.Archi ML-Enabled Systems Extension

Release: 2026-05-03

Modified on: 2026-05-03
This version:
https://meta.linked.archi/ml-systems/onto/0.1.0#
Revision:
0.1.0
Authors:
Kalin Maldzhanski
Publisher:
Linked.Archi
Source:
https://arxiv.org/abs/2308.05239
See also:
https://arxiv.org/abs/2308.05239
https://doi.org/10.1109/WAIN52551.2021.00026
https://ml-ops.org/
Download serialization:
JSON-LD RDF/XML N-Triples TTL
License:
http://creativecommons.org/licenses/by/4.0/
Visualization:
Visualize with WebVowl
Cite as:
Linked.Archi ML-Enabled Systems Extension
Provenance of this page
draft

Abstract

Ontology for modeling ML-enabled system architectures within the Linked.Archi ecosystem.

Linked.Archi ML-Enabled Systems Extension: Overview back to ToC

This ontology has the following classes and properties.

Classes

Object Properties

Data Properties

Annotation Properties

Named Individuals

Linked.Archi ML-Enabled Systems Extension: Description back to ToC

Extension ontology for modeling ML-enabled system architectures. Provides element types, relationship types, stakeholders, concerns, and viewpoints for describing the ML aspects of software systems — training pipelines, model serving, feature engineering, data lineage, and the integration boundary between ML and non-ML components. Motivated by the gap identified in Moin et al. (2023): existing architecture frameworks lack stakeholders, viewpoints, and model kinds for data scientists, data engineers, and ML engineers. This extension addresses that gap within the Linked.Archi ecosystem. Trademarked names referenced herein are property of their respective owners.

Cross-reference for Linked.Archi ML-Enabled Systems Extension classes, object properties and data properties back to ToC

This section provides details for each class and property defined by Linked.Archi ML-Enabled Systems Extension.

Classes

Data Pipelinec back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#DataPipeline

An automated workflow for ingesting, transforming, validating, and delivering data to downstream consumers (training pipelines, feature stores, analytics systems). Distinct from TrainingPipeline in that it handles raw data processing rather than model training.
has super-classes
ML Component c

Datasetc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#Dataset

A versioned collection of data used for training, validation, or evaluation of ML models. Carries metadata about provenance, schema, quality metrics, and applicable data governance policies.
has super-classes
ML Component c
is in domain of
has data lineage op

Experiment Trackerc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#ExperimentTracker

A component for tracking ML experiments — recording hyperparameters, metrics, artifacts, and code versions for each training run. Enables reproducibility and comparison across experiments.
has super-classes
ML Component c

Feature Storec back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#FeatureStore

A centralized repository for storing, managing, and serving ML features. Provides consistent feature definitions across training and serving, feature versioning, and point-in-time correctness for training data.
has super-classes
ML Component c

Integrates Withc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#IntegratesWith

A qualified integration relationship between ML and non-ML components. Carries metadata about the integration contract: API format, latency SLA, data schema, and responsible team.
has super-classes
Qualified Relationship c
is in range of
qualified Integrates With op

ML Componentc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLComponent

An abstract base class for all ML-specific architectural elements. ML components are software artifacts with special characteristics: their behavior is inferred from data rather than specified at design time, they require retraining, and they have distinct lifecycle and versioning needs.
has super-classes
Element c
has sub-classes
Data Pipeline c, Dataset c, Experiment Tracker c, Feature Store c, ML Model c, ML Monitoring Component c, Model Registry c, Serving Infrastructure c, Training Pipeline c

ML Modelc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLModel

A trained machine learning model — the core artifact produced by the data science workflow. Includes the model architecture (e.g., neural network topology, decision tree structure), trained parameters, and associated metadata (training date, dataset version, performance metrics).
has super-classes
ML Component c
is in domain of
has framework dp, has model architecture dp, has model version dp, has monitoring plan op, has performance metric dp, has retraining schedule dp

ML Monitoring Componentc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#MonitoringComponent

A component responsible for monitoring ML model behavior in production — tracking prediction distributions, detecting data drift and concept drift, alerting on performance degradation, and triggering retraining workflows.
has super-classes
ML Component c
is in range of
has monitoring plan op

Model Registryc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#ModelRegistry

A versioned catalog of trained ML models with their metadata, lineage, performance metrics, and deployment status. The single source of truth for which models are approved for production.
has super-classes
ML Component c

Produces Modelc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#ProducesModel

A qualified model production relationship — carries pipeline run metadata.
has super-classes
Qualified Relationship c
is in range of
qualified Produces Model op

Serving Infrastructurec back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#ServingInfrastructure

Infrastructure for serving ML model predictions at runtime — model servers (TensorFlow Serving, Triton, Seldon), API endpoints, batch prediction jobs, and edge deployment runtimes.
has super-classes
ML Component c
is in domain of
has serving latency SLA dp

Trained Onc back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#TrainedOn

A qualified training relationship — carries metadata about training run, date, hyperparameters.
has super-classes
Qualified Relationship c
is in range of
qualified Trained On op

Training Pipelinec back to ToC or Class ToC

IRI: https://meta.linked.archi/ml-systems/onto#TrainingPipeline

The end-to-end automated process for training an ML model — from data ingestion and preprocessing through feature engineering, model training, evaluation, and model registration. Typically orchestrated by tools like Kubeflow, MLflow, or Airflow.
has super-classes
ML Component c

Object Properties

consumes dataop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#consumesData

Links a data pipeline to the data source it ingests from.

feeds featuresop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#feedsFeatures

Links a feature store to the training pipeline or serving infrastructure it provides features to.

has data lineageop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#hasDataLineage

Links a dataset to its upstream data sources for provenance tracking.
has domain
Dataset c
has range
Element c

has monitoring planop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#hasMonitoringPlan

Links an ML model to the monitoring component responsible for tracking its production behavior.
has domain
ML Model c
has range
ML Monitoring Component c

integrates withop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#integratesWith

Links an ML component to a non-ML component it integrates with. This is the critical boundary identified by Moin et al. — the interface between ML and traditional software components.

monitorsop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#monitors

Links a monitoring component to the ML component it monitors.

produces datasetop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#producesDataset

Links a data pipeline to the dataset it produces.

produces modelop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#producesModel

Links a training pipeline to the ML model it produces.

qualified Integrates Withop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#qualifiedIntegratesWith

has range
Integrates With c

qualified Produces Modelop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#qualifiedProducesModel

has range
Produces Model c

qualified Trained Onop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#qualifiedTrainedOn

has range
Trained On c

registered inop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#registeredIn

Links an ML model to the model registry where it is cataloged.

servesop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#serves

Links serving infrastructure to the ML model it serves for inference.

trained onop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#trainedOn

Links an ML model to the dataset it was trained on.

triggers retrainingop back to ToC or Object Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#triggersRetraining

Links a monitoring component to the training pipeline it triggers when drift is detected.

Data Properties

has frameworkdp back to ToC or Data Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#hasFramework

The ML framework used to build the model (e.g., TensorFlow, PyTorch, scikit-learn).
has domain
ML Model c
has range
string

has model architecturedp back to ToC or Data Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#hasModelArchitecture

Description of the ML model architecture — e.g., 'ResNet-50', '3-layer LSTM', 'XGBoost with 500 trees'. This is the structural specification that data scientists use to communicate model design.
has domain
ML Model c
has range
string

has model versiondp back to ToC or Data Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#hasModelVersion

The version identifier of a trained ML model.
has domain
ML Model c
has range
string

has performance metricdp back to ToC or Data Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#hasPerformanceMetric

A key performance metric value for an ML model (e.g., accuracy=0.94).
has domain
ML Model c
has range
string

has retraining scheduledp back to ToC or Data Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#hasRetrainingSchedule

The schedule or trigger for retraining the model (e.g., 'weekly', 'on drift detection').
has domain
ML Model c
has range
string

has serving latency SLAdp back to ToC or Data Property ToC

IRI: https://meta.linked.archi/ml-systems/onto#hasServingLatencySLA

The maximum acceptable latency for model inference (e.g., 'p99 < 100ms').
has domain
Serving Infrastructure c
has range
string

Annotation Properties

abstractap back to ToC or Annotation Property ToC

IRI: http://purl.org/dc/terms/abstract

citationap back to ToC or Annotation Property ToC

IRI: http://schema.org/citation

createdap back to ToC or Annotation Property ToC

IRI: http://purl.org/dc/terms/created

creatorap back to ToC or Annotation Property ToC

IRI: http://purl.org/dc/terms/creator

definitionap back to ToC or Annotation Property ToC

IRI: http://www.w3.org/2004/02/skos/core#definition

descriptionap back to ToC or Annotation Property ToC

IRI: http://purl.org/dc/terms/description

domain Includesap back to ToC or Annotation Property ToC

IRI: https://meta.linked.archi/core#domainIncludes

includes Conceptap back to ToC or Annotation Property ToC

IRI: https://meta.linked.archi/core#includesConcept

licenseap back to ToC or Annotation Property ToC

IRI: http://creativecommons.org/ns#license

modifiedap back to ToC or Annotation Property ToC

IRI: http://purl.org/dc/terms/modified

pref Labelap back to ToC or Annotation Property ToC

IRI: http://www.w3.org/2004/02/skos/core#prefLabel

preferred Namespace Prefixap back to ToC or Annotation Property ToC

IRI: http://purl.org/vocab/vann/preferredNamespacePrefix

preferred Namespace Uriap back to ToC or Annotation Property ToC

IRI: http://purl.org/vocab/vann/preferredNamespaceUri

publisherap back to ToC or Annotation Property ToC

IRI: http://purl.org/dc/terms/publisher

range Includesap back to ToC or Annotation Property ToC

IRI: https://meta.linked.archi/core#rangeIncludes

scope Noteap back to ToC or Annotation Property ToC

IRI: http://www.w3.org/2004/02/skos/core#scopeNote

sourceap back to ToC or Annotation Property ToC

IRI: http://purl.org/dc/terms/source

statusap back to ToC or Annotation Property ToC

IRI: http://purl.org/ontology/bibo/status

targets Stakeholderap back to ToC or Annotation Property ToC

IRI: https://meta.linked.archi/core#targetsStakeholder

titleap back to ToC or Annotation Property ToC

IRI: http://purl.org/dc/terms/title

unqualified Formap back to ToC or Annotation Property ToC

IRI: https://meta.linked.archi/core#unqualifiedForm

view Typeap back to ToC or Annotation Property ToC

IRI: https://meta.linked.archi/core#viewType

viewpoint Frames Concernap back to ToC or Annotation Property ToC

IRI: https://meta.linked.archi/core#viewpointFramesConcern

viewpoint Has Purposeap back to ToC or Annotation Property ToC

IRI: https://meta.linked.archi/core#viewpointHasPurpose

was Derived Fromap back to ToC or Annotation Property ToC

IRI: http://www.w3.org/ns/prov#wasDerivedFrom

Named Individuals

Adversarial Robustnessni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#AdversarialRobustnessConcern

Resilience of ML models against adversarial attacks — deliberate manipulation of inputs or training data to cause incorrect predictions. Includes evasion attacks, data poisoning, and model extraction.
belongs to
Concern c
has facts
definition ap "Resilience of ML models against adversarial attacks — deliberate manipulation of inputs or training data to cause incorrect predictions. Includes evasion attacks, data poisoning, and model extraction."@en
pref Label ap "Adversarial Robustness"@en
scope Note ap "Moin et al. (2023) note that software engineers often underestimate the vulnerability of ML-enabled systems through adversarial attacks."@en

Analytics Modelingni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#AnalyticsModelingViewpoint

Frames the concerns of data scientists — model architecture, training pipelines, feature engineering, evaluation metrics, and algorithm selection. Views governed by this viewpoint show the ML model landscape: which models exist, what data they consume, how they are trained, and how they perform.
belongs to
Viewpoint c
has facts
definition ap "Frames the concerns of data scientists — model architecture, training pipelines, feature engineering, evaluation metrics, and algorithm selection. Views governed by this viewpoint show the ML model landscape: which models exist, what data they consume, how they are trained, and how they perform."@en
pref Label ap "Analytics Modeling"@en
includes Concept ap Dataset c
includes Concept ap Experiment Tracker c
includes Concept ap Feature Store c
includes Concept ap ML Model c
includes Concept ap Training Pipeline c
targets Stakeholder ap Data Scientist ni
targets Stakeholder ap ML Engineer ni
view Type ap Diagram ep
viewpoint Frames Concern ap Data Quality ni
viewpoint Frames Concern ap Explainability ni
viewpoint Frames Concern ap Fairness & Bias ni
viewpoint Frames Concern ap ML Model Performance ni
viewpoint Has Purpose ap Designing ep

Analytics Operationsni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#AnalyticsOperationsViewpoint

Frames the concerns of data engineers and ML engineers — model deployment, serving infrastructure, monitoring, retraining pipelines, and operational health. Views governed by this viewpoint show the MLOps landscape: how models move from training to production and how they are maintained.
belongs to
Viewpoint c
has facts
definition ap "Frames the concerns of data engineers and ML engineers — model deployment, serving infrastructure, monitoring, retraining pipelines, and operational health. Views governed by this viewpoint show the MLOps landscape: how models move from training to production and how they are maintained."@en
pref Label ap "Analytics Operations"@en
includes Concept ap Data Pipeline c
includes Concept ap Model Registry c
includes Concept ap ML Monitoring Component c
includes Concept ap Serving Infrastructure c
includes Concept ap Training Pipeline c
targets Stakeholder ap Data Engineer ni
targets Stakeholder ap ML Engineer ni
view Type ap Diagram ep
viewpoint Frames Concern ap ML Artifact Versioning ni
viewpoint Frames Concern ap ML Monitorability ni
viewpoint Frames Concern ap Model Drift ni
viewpoint Has Purpose ap Designing ep
viewpoint Has Purpose ap Governing ep

Data Engineerni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#DataEngineer

Responsible for analytics operations — building and maintaining data pipelines, ensuring scalable data processing, deploying trained models into production, and managing the infrastructure for model serving and monitoring.
belongs to
Stakeholder c
has facts
definition ap "Responsible for analytics operations — building and maintaining data pipelines, ensuring scalable data processing, deploying trained models into production, and managing the infrastructure for model serving and monitoring."@en
pref Label ap "Data Engineer"@en
scope Note ap "Identified by Moin et al. (2023) as a missing stakeholder. Their concerns include data pipeline reliability, model deployment automation, serving latency, and infrastructure scalability."@en

Data Qualityni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#DataQualityConcern

Completeness, correctness, timeliness, representativeness, and consistency of data used for training, validation, and inference. Poor data quality directly degrades model performance and can introduce bias.
belongs to
Concern c
has facts
definition ap "Completeness, correctness, timeliness, representativeness, and consistency of data used for training, validation, and inference. Poor data quality directly degrades model performance and can introduce bias."@en
pref Label ap "Data Quality"@en

Data Scientistni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#DataScientist

Responsible for analytics modeling — selecting algorithms, engineering features, training ML models, and evaluating model performance using metrics such as accuracy, precision, recall, and F1 score. Data scientists work primarily with mathematical formalisms, statistical models, and computational graphs.
belongs to
Stakeholder c
has facts
definition ap "Responsible for analytics modeling — selecting algorithms, engineering features, training ML models, and evaluating model performance using metrics such as accuracy, precision, recall, and F1 score. Data scientists work primarily with mathematical formalisms, statistical models, and computational graphs."@en
pref Label ap "Data Scientist"@en
scope Note ap "Identified by Moin et al. (2023) as a missing stakeholder in existing architecture frameworks. Their concerns include ML model accuracy, feature quality, algorithm selection, and model interpretability."@en

Ethics Committee / Officerni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#EthicsOfficer

Responsible for evaluating the ethical, legal, and social implications (ELSI) of ML-enabled systems — including fairness, bias, explainability, privacy, and trustworthiness of ML components.
belongs to
Stakeholder c
has facts
definition ap "Responsible for evaluating the ethical, legal, and social implications (ELSI) of ML-enabled systems — including fairness, bias, explainability, privacy, and trustworthiness of ML components."@en
pref Label ap "Ethics Committee / Officer"@en
scope Note ap "Identified by Moin et al. (2023) as a stakeholder whose concerns are not addressed by existing architecture frameworks."@en

Explainabilityni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#ExplainabilityConcern

The ability to understand, interpret, and communicate why an ML model produces specific predictions. Critical for regulatory compliance (e.g., GDPR right to explanation), stakeholder trust, and debugging.
belongs to
Concern c
has facts
definition ap "The ability to understand, interpret, and communicate why an ML model produces specific predictions. Critical for regulatory compliance (e.g., GDPR right to explanation), stakeholder trust, and debugging."@en
pref Label ap "Explainability"@en

Fairness & Biasni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#FairnessConcern

Ensuring ML models do not systematically discriminate against individuals or groups based on protected attributes (race, gender, age, etc.). Includes detection and mitigation of bias in training data and model outputs.
belongs to
Concern c
has facts
definition ap "Ensuring ML models do not systematically discriminate against individuals or groups based on protected attributes (race, gender, age, etc.). Includes detection and mitigation of bias in training data and model outputs."@en
pref Label ap "Fairness & Bias"@en

ML Artifact Versioningni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLArtifactVersioningConcern

Tracking versions of ML artifacts — trained models, model parameters and hyperparameters, training datasets, feature definitions, and pipeline configurations. Essential for reproducibility and auditability.
belongs to
Concern c
has facts
definition ap "Tracking versions of ML artifacts — trained models, model parameters and hyperparameters, training datasets, feature definitions, and pipeline configurations. Essential for reproducibility and auditability."@en
pref Label ap "ML Artifact Versioning"@en

ML Engineerni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLEngineer

Bridges data science and software engineering — productionizing ML models, managing model serving infrastructure, implementing monitoring and retraining pipelines, and ensuring ML components integrate correctly with non-ML system components.
Source
https://doi.org/10.1109/WAIN52551.2021.00026
belongs to
Stakeholder c
has facts
source ap W A I N52551.2021.00026 ep
definition ap "Bridges data science and software engineering — productionizing ML models, managing model serving infrastructure, implementing monitoring and retraining pipelines, and ensuring ML components integrate correctly with non-ML system components."@en
pref Label ap "ML Engineer"@en
scope Note ap "Recognized by Muccini & Vaidhyanathan (2021) and Lewis et al. (2021) as a distinct role in ML-enabled system development."@en

ML Ethics & Complianceni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLEthicsViewpoint

Frames fairness, explainability, privacy, and regulatory compliance concerns for ML-enabled systems. Views governed by this viewpoint show which models handle sensitive data, what bias mitigation measures are in place, and compliance status with applicable regulations.
belongs to
Viewpoint c
has facts
definition ap "Frames fairness, explainability, privacy, and regulatory compliance concerns for ML-enabled systems. Views governed by this viewpoint show which models handle sensitive data, what bias mitigation measures are in place, and compliance status with applicable regulations."@en
pref Label ap "ML Ethics & Compliance"@en
includes Concept ap Dataset c
includes Concept ap ML Model c
targets Stakeholder ap Ethics Committee / Officer ni
view Type ap Catalog ep
view Type ap Matrix ep
viewpoint Frames Concern ap Adversarial Robustness ni
viewpoint Frames Concern ap Explainability ni
viewpoint Frames Concern ap Fairness & Bias ni
viewpoint Frames Concern ap ML Privacy ni
viewpoint Has Purpose ap Governing ep

ML Model Catalogni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLModelCatalogViewpoint

Inventory of all ML models with key attributes — owner, status, performance metrics, dataset lineage, deployment target, and monitoring status. The ML equivalent of a component catalog.
belongs to
Viewpoint c
has facts
definition ap "Inventory of all ML models with key attributes — owner, status, performance metrics, dataset lineage, deployment target, and monitoring status. The ML equivalent of a component catalog."@en
pref Label ap "ML Model Catalog"@en
includes Concept ap ML Model c
includes Concept ap Model Registry c
targets Stakeholder ap Data Engineer ni
targets Stakeholder ap Data Scientist ni
targets Stakeholder ap ML Engineer ni
view Type ap Catalog ep
viewpoint Frames Concern ap ML Artifact Versioning ni
viewpoint Frames Concern ap ML Model Performance ni
viewpoint Has Purpose ap Governing ep
viewpoint Has Purpose ap Informing ep

ML Model Performanceni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#ModelPerformanceConcern

ML-specific performance metrics — accuracy, precision, recall, F1 score, AUC-ROC, mean squared error, and other evaluation metrics that measure how well an ML model performs its inference task on unseen data.
belongs to
Concern c
has facts
definition ap "ML-specific performance metrics — accuracy, precision, recall, F1 score, AUC-ROC, mean squared error, and other evaluation metrics that measure how well an ML model performs its inference task on unseen data."@en
pref Label ap "ML Model Performance"@en
scope Note ap "Distinct from system performance (latency, throughput). A model may have excellent ML performance but introduce unacceptable latency, or vice versa."@en

ML Monitorabilityni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLMonitorabilityConcern

Runtime monitoring of ML components — prediction distributions, feature distributions, serving latency, model staleness, and data pipeline health. Enables detection of drift, degradation, and anomalies.
belongs to
Concern c
has facts
definition ap "Runtime monitoring of ML components — prediction distributions, feature distributions, serving latency, model staleness, and data pipeline health. Enables detection of drift, degradation, and anomalies."@en
pref Label ap "ML Monitorability"@en
scope Note ap "Lewis et al. (2021) identify monitorability as a new quality attribute concern specific to ML-enabled systems."@en

ML Privacyni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLPrivacyConcern

Privacy challenges specific to ML — training data memorization, model inversion attacks, membership inference, and compliance with data protection regulations (GDPR, CCPA) in the context of ML pipelines.
belongs to
Concern c
has facts
definition ap "Privacy challenges specific to ML — training data memorization, model inversion attacks, membership inference, and compliance with data protection regulations (GDPR, CCPA) in the context of ML pipelines."@en
pref Label ap "ML Privacy"@en

ML System Integrationni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLSystemIntegrationViewpoint

Shows how ML components integrate with non-ML components — the boundary that Moin et al. identify as the key collaboration challenge. Views governed by this viewpoint show integration contracts, data flows, latency budgets, and team responsibilities at the ML/SE boundary.
belongs to
Viewpoint c
has facts
definition ap "Shows how ML components integrate with non-ML components — the boundary that Moin et al. identify as the key collaboration challenge. Views governed by this viewpoint show integration contracts, data flows, latency budgets, and team responsibilities at the ML/SE boundary."@en
pref Label ap "ML System Integration"@en
includes Concept ap Integrates With c
includes Concept ap ML Component c
targets Stakeholder ap Data Engineer ni
targets Stakeholder ap Data Scientist ni
targets Stakeholder ap ML Engineer ni
view Type ap Diagram ep
view Type ap Matrix ep
viewpoint Frames Concern ap ML–SE Collaboration ni
viewpoint Has Purpose ap Designing ep
viewpoint Has Purpose ap Informing ep

ML–SE Collaborationni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#MLSECollaborationConcern

The challenge of coordinating work between data science and software engineering teams — different vocabularies, tools, workflows, and quality metrics. The "ML Mismatch" problem identified by Lewis et al. (2021).
Source
https://doi.org/10.1109/WAIN52551.2021.00026
belongs to
Concern c
has facts
source ap W A I N52551.2021.00026 ep
definition ap "The challenge of coordinating work between data science and software engineering teams — different vocabularies, tools, workflows, and quality metrics. The "ML Mismatch" problem identified by Lewis et al. (2021)."@en
pref Label ap "ML–SE Collaboration"@en

Model Driftni back to ToC or Named Individual ToC

IRI: https://meta.linked.archi/ml-systems/onto#ModelDriftConcern

Degradation of model performance over time due to changes in the underlying data distribution (data drift) or changes in the relationship between features and targets (concept drift). Requires monitoring and retraining strategies.
belongs to
Concern c
has facts
definition ap "Degradation of model performance over time due to changes in the underlying data distribution (data drift) or changes in the relationship between features and targets (concept drift). Requires monitoring and retraining strategies."@en
pref Label ap "Model Drift"@en

Legend back to ToC

c: Classes
op: Object Properties
dp: Data Properties
ni: Named Individuals
ep: External Properties

Acknowledgments back to ToC

The authors would like to thank Silvio Peroni for developing LODE, a Live OWL Documentation Environment, which is used for representing the Cross Referencing Section of this document and Daniel Garijo for developing Widoco, the program used to create the template used in this documentation.