Acronyms

Summary

Acronyms are short forms that recur across the wiki. Use this page when reading a page quickly and follow the linked concept page when the acronym names a method or metric that needs more detail.

Common acronyms

AcronymExpansionArea
ALSAlternating least squaresRecommendation systems
ANNApproximate nearest-neighbour searchSearch
ARIMAAutoregressive integrated moving averageForecasting
BM25Best Matching 25 retrieval functionSearch
CNNConvolutional neural networkDeep learning
DQNDeep Q-networkReinforcement learning
ETL / ELTExtract-transform-load / extract-load-transformData engineering
GANGenerative adversarial networkDeep learning
GBMGradient-boosted machine (trees)Classical ML
GRUGated recurrent unitDeep learning
JEPAJoint Embedding Predictive ArchitectureVideo understanding
LLMLarge language modelGenerative AI
LSTMLong short-term memoryDeep learning
MAPMaximum a posteriori (estimation)Probability
MDPMarkov decision processReinforcement learning
MLEMaximum likelihood estimationProbability
NDCGNormalized discounted cumulative gainSearch
NERNamed entity recognitionNLP
OCROptical character recognitionNLP
PCAPrincipal component analysisClassical ML
RAGRetrieval-augmented generationGenerative AI
RNNRecurrent neural networkDeep learning
SLOService-level objectiveMLOps
SVDSingular value decompositionMathematical foundations
SVMSupport vector machineClassical ML
TF-IDFTerm frequency-inverse document frequencySearch
VAEVariational autoencoderDeep learning
VLMVision-language modelGenerative AI

Retrieval metric acronyms

The retrieval metrics mean reciprocal rank and normalized discounted cumulative gain are linked from the metric reference page because their short forms are validated elsewhere in the corpus.

Usage guidance

  • Expand an acronym on first use in long-form pages.
  • Prefer the canonical concept title in headings.
  • Link to glossary, notation, or the canonical concept page when the short form alone is ambiguous.
  • Avoid inventing project-specific acronyms unless the page defines them.