# Model Architecture > Every term in this category is listed as one record, with its kind, its definition and its aliases, the principles whose relations name it, the principle or… Page: Ontology · Lexicon Canonical: https://banes-lab.com/ontology/lexicon#lex-category-model-architecture Every term in this category is listed as one record, with its kind, its definition and its aliases, the principles whose relations name it, the principle or contract that carries the same name where one exists, and the layer its category belongs to. ### Ad-Hoc Notebook-to-Production - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Promoting exploratory notebook code straight to production without engineering it into a reliable pipeline. Referenced by [Machine Learning Architecture](https://banes-lab.com/records/arch/machine-learning-architecture.md) ### Approval Policy - Kind: [constraint](https://banes-lab.com/records/kind/constraint.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The declared rules and gates a model must pass before it may be deployed. Referenced by [Model Governance](https://banes-lab.com/records/arch/model-governance.md) ### Audit and Debugging - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to inspect and trace a model's decisions for auditing and debugging. Referenced by [Explainability](https://banes-lab.com/records/arch/explainability.md) ### Bounded Tool-Using Agents - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to run agents that use external tools within defined, safe limits. Referenced by [Agentic Architecture](https://banes-lab.com/records/arch/agentic-architecture.md) ### Capability/Utility - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The degree of usefulness a model offers, which strict safety limits can constrain. Referenced by [Model Safety](https://banes-lab.com/records/arch/model-safety.md) ### Contextual Generation - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to generate output informed by retrieved, task-specific context. Referenced by [Retrieval-Augmented Generation (RAG)](https://banes-lab.com/records/arch/retrieval-augmented-generation.md) ### Controlled Model Deployment - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to release models through a governed, approved process. Referenced by [Model Governance](https://banes-lab.com/records/arch/model-governance.md) ### Curation Cost - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Performance Core](https://banes-lab.com/records/layer/performance-core.md) Details Definition The degree of ongoing effort required to build and maintain a curated knowledge graph. Referenced by [Knowledge Graphs](https://banes-lab.com/records/arch/knowledge-graphs.md) ### Data Pipeline - Kind: [mechanism](https://banes-lab.com/records/kind/mechanism.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The stages that ingest, clean, and transform data into a form suitable for training or inference. Referenced by [Machine Learning Architecture](https://banes-lab.com/records/arch/machine-learning-architecture.md) ### Data/Model Boundaries - Kind: [constraint](https://banes-lab.com/records/kind/constraint.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The lines separating data preparation, model training, and serving so each concern stays isolated. Referenced by [Artificial Intelligence Architecture](https://banes-lab.com/records/arch/artificial-intelligence-architecture.md) ### Dataset - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition A curated collection of examples used to train or evaluate a model. Referenced by [Model Evaluation](https://banes-lab.com/records/arch/model-evaluation.md) ### Degradation Detection - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to detect when a model's accuracy declines as data shifts. Referenced by [Model Drift Monitoring](https://banes-lab.com/records/arch/model-drift-monitoring.md) ### Deploy-and-Forget Models - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Deploying a model and never monitoring it, so degradation as the data drifts goes unnoticed. Referenced by [Model Drift Monitoring](https://banes-lab.com/records/arch/model-drift-monitoring.md) ### Document Store - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The body of documents a retrieval system searches to ground a model's generation. Referenced by [Retrieval-Augmented Generation (RAG)](https://banes-lab.com/records/arch/retrieval-augmented-generation.md) ### Embeddings - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Numeric vector representations of data that place semantically similar items near each other. Referenced by [Vector Search](https://banes-lab.com/records/arch/vector-search.md) ### Entities - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The distinct things, such as people, places or concepts, that a knowledge graph represents as nodes. Referenced by [Knowledge Graphs](https://banes-lab.com/records/arch/knowledge-graphs.md) ### Exact Keyword Search Only - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Relying solely on exact keyword matching for retrieval, missing semantically related results. Referenced by [Vector Search](https://banes-lab.com/records/arch/vector-search.md) ### Experiment Velocity - Kind: [metric](https://banes-lab.com/records/kind/metric.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The rate at which model experiments can be run and iterated, which governance can slow. Referenced by [Model Governance](https://banes-lab.com/records/arch/model-governance.md) ### Experimentation Speed - Kind: [metric](https://banes-lab.com/records/kind/metric.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The rate at which new modeling ideas can be tried and evaluated. Referenced by [Machine Learning Architecture](https://banes-lab.com/records/arch/machine-learning-architecture.md) ### Explainability/Recall - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The degree to which retrieval stays explainable and complete, traded against pure similarity ranking. Referenced by [Vector Search](https://banes-lab.com/records/arch/vector-search.md) ### Flat Document-Only Knowledge - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Representing knowledge as unlinked flat documents, losing the relationships a graph would capture. Referenced by [Knowledge Graphs](https://banes-lab.com/records/arch/knowledge-graphs.md) ### Governed Autonomy - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to let an agent act autonomously within enforced governance limits. Referenced by [Agentic Architecture](https://banes-lab.com/records/arch/agentic-architecture.md) ### Grounding Strategy - Kind: [approach](https://banes-lab.com/records/kind/approach.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition A scheme for anchoring a model's output in retrieved, authoritative sources rather than its parameters alone. Referenced by [Retrieval-Augmented Generation (RAG)](https://banes-lab.com/records/arch/retrieval-augmented-generation.md) ### Guardrails - Kind: [mechanism](https://banes-lab.com/records/kind/mechanism.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Constraints and filters that bound what a model is permitted to output or do at runtime. Referenced by [Model Safety](https://banes-lab.com/records/arch/model-safety.md) ### Input/Output Contract - Kind: [constraint](https://banes-lab.com/records/kind/constraint.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The agreed schema of the inputs a model accepts and the outputs it returns. Referenced by [Model Inference](https://banes-lab.com/records/arch/model-inference.md) ### Knowledge Freshness - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The degree to which a system's knowledge reflects current rather than stale information. Referenced by [Retrieval-Augmented Generation (RAG)](https://banes-lab.com/records/arch/retrieval-augmented-generation.md) ### Latency/Cost - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Performance Core](https://banes-lab.com/records/layer/performance-core.md) Details Definition The degree of latency and expense incurred to serve model predictions. Referenced by [Model Inference](https://banes-lab.com/records/arch/model-inference.md) ### Metric Completeness - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The degree to which evaluation metrics capture every dimension of a model's quality. Referenced by [Model Evaluation](https://banes-lab.com/records/arch/model-evaluation.md) ### Model Artifact - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The trained model file, with its learned weights, that is loaded to serve predictions. Referenced by [Model Inference](https://banes-lab.com/records/arch/model-inference.md) ### Model Complexity - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The degree of intricacy in a model, which raises accuracy but lowers explainability. Referenced by [Explainability](https://banes-lab.com/records/arch/explainability.md) ### Model Registry - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition A catalog that tracks model versions, their metadata, and their deployment status. Referenced by [Model Governance](https://banes-lab.com/records/arch/model-governance.md) ### Model Selection/Regression Detection - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to compare models and catch quality regressions before deployment. Referenced by [Model Evaluation](https://banes-lab.com/records/arch/model-evaluation.md) ### Model-Integrated Systems - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability of a software system to incorporate models as integral parts of its behavior. Referenced by [Artificial Intelligence Architecture](https://banes-lab.com/records/arch/artificial-intelligence-architecture.md) ### Monitoring Cost - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Performance Core](https://banes-lab.com/records/layer/performance-core.md) Details Definition The degree of ongoing expense of continuously monitoring a deployed model. Referenced by [Model Drift Monitoring](https://banes-lab.com/records/arch/model-drift-monitoring.md) ### Opaque Black-Box Decisions - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Producing model decisions with no explanation, so their reasoning cannot be inspected or trusted. Referenced by [Explainability](https://banes-lab.com/records/arch/explainability.md) ### Opaque Ungoverned Model Use - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Using models with no governance or oversight, leaving their behavior and risks unmanaged. Referenced by [Artificial Intelligence Architecture](https://banes-lab.com/records/arch/artificial-intelligence-architecture.md) ### Rationale/Evidence - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The reasons and supporting evidence recorded for a model's decision. Referenced by [Explainability](https://banes-lab.com/records/arch/explainability.md) ### Relations - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The typed connections between entities that a knowledge graph represents as edges. Referenced by [Knowledge Graphs](https://banes-lab.com/records/arch/knowledge-graphs.md) ### Relationship-Aware Retrieval/Reasoning - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to retrieve and reason over the relationships between entities, not just isolated facts. Referenced by [Knowledge Graphs](https://banes-lab.com/records/arch/knowledge-graphs.md) ### Reliable Model Lifecycle - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to manage a model's data, training, deployment and monitoring reliably and repeatably. Referenced by [Machine Learning Architecture](https://banes-lab.com/records/arch/machine-learning-architecture.md) ### Retraining Triggers - Kind: [mechanism](https://banes-lab.com/records/kind/mechanism.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Signals that automatically initiate model retraining when measured drift crosses a threshold. Referenced by [Model Drift Monitoring](https://banes-lab.com/records/arch/model-drift-monitoring.md) ### Retrieval Quality/Latency - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Performance Core](https://banes-lab.com/records/layer/performance-core.md) Details Definition The degree to which retrieval must trade result quality against speed. Referenced by [Retrieval-Augmented Generation (RAG)](https://banes-lab.com/records/arch/retrieval-augmented-generation.md) ### Retriever - Kind: [mechanism](https://banes-lab.com/records/kind/mechanism.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition A component that finds and returns the most relevant documents for a query. Referenced by [Retrieval-Augmented Generation (RAG)](https://banes-lab.com/records/arch/retrieval-augmented-generation.md) ### Runtime Prediction/Generation - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to produce predictions or generated output from a trained model at runtime. Referenced by [Model Inference](https://banes-lab.com/records/arch/model-inference.md) ### Safe Model Deployment - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to deploy models with safeguards that bound their behavior. Referenced by [Model Safety](https://banes-lab.com/records/arch/model-safety.md) ### Schema/Ontology - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition A formal definition of the entity types and relationship types a knowledge graph may contain. Referenced by [Knowledge Graphs](https://banes-lab.com/records/arch/knowledge-graphs.md) ### Semantic Search - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to find results by meaning and similarity rather than exact keyword match. Referenced by [Vector Search](https://banes-lab.com/records/arch/vector-search.md) ### Similarity Retrieval - Kind: [capability](https://banes-lab.com/records/kind/capability.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The ability to retrieve items nearest to a query in an embedding space. Referenced by [Vector Search](https://banes-lab.com/records/arch/vector-search.md) ### Structured, Versioned Prompts - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Prompts authored as structured, version-controlled artifacts rather than ad-hoc strings. Referenced by [Prompt Engineering](https://banes-lab.com/records/arch/prompt-engineering.md) ### Tool Interface - Kind: [constraint](https://banes-lab.com/records/kind/constraint.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The defined contract through which an agent invokes external tools and receives their results. Referenced by [Agentic Architecture](https://banes-lab.com/records/arch/agentic-architecture.md) ### Training-Time-Only Model Logic - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Building logic that exists only during training, with no counterpart to serve predictions at inference. Referenced by [Model Inference](https://banes-lab.com/records/arch/model-inference.md) ### Training/Inference Separation - Kind: [constraint](https://banes-lab.com/records/kind/constraint.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The requirement that model training and prediction serving be distinct, separately-managed phases. Referenced by [Machine Learning Architecture](https://banes-lab.com/records/arch/machine-learning-architecture.md) ### Trust - Kind: [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition The degree to which users are willing to rely on a system's outputs. Referenced by [Explainability](https://banes-lab.com/records/arch/explainability.md) ### Unapproved Model Deployment - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Deploying a model to production without passing the required review and approval gates. Referenced by [Model Governance](https://banes-lab.com/records/arch/model-governance.md) ### Ungrounded Generation - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Generating output from a model alone without grounding it in retrieved facts, inviting hallucination. Referenced by [Retrieval-Augmented Generation (RAG)](https://banes-lab.com/records/arch/retrieval-augmented-generation.md) ### Unguarded Model Autonomy - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Letting a model act autonomously with no safety guardrails on what it can do. Referenced by [Model Safety](https://banes-lab.com/records/arch/model-safety.md) ### Untested Model Deployment - Kind: [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition Deploying a model without evaluating it, so its real-world quality is unknown until it fails. Referenced by [Model Evaluation](https://banes-lab.com/records/arch/model-evaluation.md) ### Vector Index - Kind: [artifact](https://banes-lab.com/records/kind/artifact.md) - Category: [Model Architecture](https://banes-lab.com/ontology/lexicon/lex-category-model-architecture.md) - Layer: [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) Details Definition A data structure that organizes embedding vectors for fast nearest-neighbor lookup. Referenced by [Vector Search](https://banes-lab.com/records/arch/vector-search.md) ## Links to - [anti-pattern](https://banes-lab.com/records/kind/anti-pattern.md) - [Correctness Core](https://banes-lab.com/records/layer/correctness-core.md) - [Machine Learning Architecture](https://banes-lab.com/records/arch/machine-learning-architecture.md) - [constraint](https://banes-lab.com/records/kind/constraint.md) - [Model Governance](https://banes-lab.com/records/arch/model-governance.md) - [capability](https://banes-lab.com/records/kind/capability.md) - [Explainability](https://banes-lab.com/records/arch/explainability.md) - [Agentic Architecture](https://banes-lab.com/records/arch/agentic-architecture.md) - [quality-attribute](https://banes-lab.com/records/kind/quality-attribute.md) - [Model Safety](https://banes-lab.com/records/arch/model-safety.md) - [Retrieval-Augmented Generation (RAG)](https://banes-lab.com/records/arch/retrieval-augmented-generation.md) - [Performance Core](https://banes-lab.com/records/layer/performance-core.md) - [Knowledge Graphs](https://banes-lab.com/records/arch/knowledge-graphs.md) - [mechanism](https://banes-lab.com/records/kind/mechanism.md) - [Artificial Intelligence Architecture](https://banes-lab.com/records/arch/artificial-intelligence-architecture.md) - [artifact](https://banes-lab.com/records/kind/artifact.md) - [Model Evaluation](https://banes-lab.com/records/arch/model-evaluation.md) - [Model Drift Monitoring](https://banes-lab.com/records/arch/model-drift-monitoring.md) - [Vector Search](https://banes-lab.com/records/arch/vector-search.md) - [metric](https://banes-lab.com/records/kind/metric.md) - [approach](https://banes-lab.com/records/kind/approach.md) - [Model Inference](https://banes-lab.com/records/arch/model-inference.md) - [Prompt Engineering](https://banes-lab.com/records/arch/prompt-engineering.md)