AI Factory Glossary

active learning for annotation,data

Select most informative examples to annotate.

active learning for inspection, data analysis

Selectively label most informative examples.

active learning,data efficiency

Iteratively select most informative examples for human labeling.

active prompting, prompting techniques

Active prompting selects most informative examples for few-shot demonstrations.

active retrieval, rag

Decide when to retrieve.

active shift, model optimization

Active shift learns optimal shift directions and magnitudes for each layer.

activity network, quality & reliability

Activity network diagrams schedule projects showing task dependencies.

ad copy,marketing,persuade

Write advertising copy. Persuasive, action-oriented.

ad creative generation,content creation

Generate advertisement visuals.

adabelief, optimization

Adaptive optimizer based on belief in gradient.

adaboost,adaptive,weight

AdaBoost adapts weights for misclassified. Early boosting method.

adafactor, optimization

Memory-efficient adaptive optimizer.

adam optimizer,model training

Adaptive learning rate optimizer using momentum and RMSprop.

adam,adamw,optimizer

AdamW is standard optimizer for LLMs. Decoupled weight decay. Momentum and adaptive learning rates.

adamw optimizer for vit, computer vision

Preferred optimizer for ViT training.

adamw,model training

Adam with decoupled weight decay regularization.

adapter layers,fine-tuning

Small trainable modules inserted into frozen pretrained models for task-specific adaptation.

adapter-based continual learning, continual learning

Add adapters for new tasks.

adaptive activation functions, neural architecture

Learn activation shapes.

adaptive aging compensation, design

Adjust operation as device ages.

adaptive attacks, ai safety

Attacks tailored to specific defense.

adaptive body bias, abb, design

Adjust body voltage to compensate variation.

adaptive body bias, design & verification

Adaptive body biasing dynamically adjusts substrate voltage compensating for PVT variations.

adaptive body biasing,design

Dynamically adjust based on conditions.

adaptive computation time (act),adaptive computation time,act,optimization

Variable computation per input.

adaptive control charts, spc

Adjust parameters over time.

adaptive discriminator augmentation (ada),adaptive discriminator augmentation,ada,generative models

Prevent overfitting in GANs.

adaptive doe, doe

Update experiments based on results.

adaptive equalization, signal & power integrity

Adaptive equalization automatically adjusts filter coefficients based on received signal characteristics.

adaptive inference, model optimization

Adaptive inference adjusts model capacity or computation based on input difficulty.

adaptive inference, optimization

Adjust computation based on input difficulty.

adaptive instance normalization in stylegan, generative models

Control style via normalization.

adaptive instance normalization, generative models

Normalize and modulate with style.

adaptive layer depth, llm architecture

Skip or repeat layers based on input complexity.

adaptive masking, nlp

Adjust masking based on difficulty.

adaptive rag, rag

Dynamically adjust retrieval strategy.

adaptive rag, rag

Adaptive RAG selects retrieval strategies based on query characteristics.

adaptive testing, testing

Adjust testing based on early results.

adaptive testing, yield enhancement

Adaptive testing modifies test content or limits based on prior results or device characteristics optimizing test time and coverage.

adaptive token selection, optimization

Select most informative tokens.

adaptive voltage scaling (avs),adaptive voltage scaling,avs,design

Dynamically adjust voltage based on workload.

adaptive voltage scaling advanced, avs, design

Dynamically adjust voltage based on chip capability.

adasyn, adasyn, machine learning

Adaptive generation of synthetic samples.

additive angular margin, audio & speech

Additive angular margin loss enhances speaker discrimination by adding angular penalties to embeddings.

additive hawkes, time series models

Additive Hawkes processes decompose intensity into baseline plus excitation from each past event.

additive noise models, time series models

Additive noise models assume effects are deterministic functions of causes plus independent noise enabling causal discovery.

adhesive bonding, advanced packaging

Bond using polymer adhesive.

adjacency matrix nas, neural architecture search

Adjacency matrix encoding represents architecture graphs as matrices for graph neural network processing.

adjoint sensitivity method, optimization

Efficient gradient computation for ODEs.

adjusted r-squared, quality & reliability

Adjusted R-squared accounts for number of predictors preventing overfitting.