MLOps Glossary
MLOps can be a daunting and confusing world, even for the most advanced minds. This is why we’ve put together a glossary of terms. You’ll find these used across the Seldon website and throughout the industry.
a
- A/B Testing
- Accountability
b
- Batch
- Beam Search
- Binary classification
- Black box model
c
- Categorical feature
- CI/CD
- Classification
- Clustering
- Covariate shift
d
- Data-centric
- Data lake
- Dataflow
- Decision rule
- Deployment
- Dimensionality reduction
- Drift detection
e
- Explainability
- Explainer
- Explanation
- Exploratory data analysis
f
- False negative
- False positive
- Feature
- Feature distribution
- Feature selection
- Feedback
- Few-shot
g
- Governance
i
- Inference
- Instance
k
- Kubernetes
l
- Labels
- Large Language Models
m
- Machine learning lifecycle:
- Metric
- Modality
- Modeling
- Multi-armed bandits
n
- Numerical feature
o
- Observability
- Offline model
- One-hot-encoding
- Online model
- Ordinal encoding
- Outlier detection
p
- Performance
- Predictions
- Probabilistic model
r
- Ranking and selection
- ReACT Agent
- Regression
- Reinforcement learning
- Request
- Runtime
s
- Scalability
- SDK
- Serving
- Supervised learning
t
- Tabular data
- Training
u
- Unsupervised learning
w
- Whitebox model
z
- Zero-shot