AI Factory Glossary

compute-constrained regime, training

Limited by computational resources.

compute-optimal scaling, training

Balance model size and data.

computer vision for wafer inspection, data analysis

Image analysis for defect detection.

concept activation, interpretability

Concept Activation Vectors identify directions in representation space corresponding to human concepts.

concept bottleneck models, explainable ai

Interpretable models using human-understandable concepts.

concept drift over time,mlops

When data distribution changes during deployment.

concept drift,mlops

When the relationship between inputs and outputs changes.

concept,tcav,high level

TCAV tests sensitivity to high-level concepts. User-defined concepts like stripes or doctor.

concolic execution,software engineering

Combine concrete and symbolic execution.

concurrency,thread,async,parallel

I explain the difference between threads, async, and processes, and help you choose patterns that fit your workload.

concurrent engineering, design

Parallel design activities.

conda environments, infrastructure

Python environment manager.

conda,environment,scientific

Conda manages environments. Scientific packages, cross-platform.

condconv, computer vision

Input-dependent convolution weights.

condition monitoring, manufacturing operations

Condition monitoring tracks equipment health through vibration temperature or performance parameters.

condition-based maintenance, production

Maintain when indicators show need.

conditional batch normalization, neural architecture

Modulate BN with conditional input.

conditional computation advanced, neural architecture

Activate network components based on input.

conditional computation efficiency, moe

Effective FLOPS with sparse activation.

conditional computation, llm architecture

Conditional computation activates different model components based on input.

conditional computation, model optimization

Conditional computation activates different network parts per input through gating or routing.

conditional computation, optimization

Activate subsets of network conditionally.

conditional control inputs, generative models

Edge maps depth poses etc.

conditional domain adaptation, domain adaptation

Align features per class.

conditional graph gen, graph neural networks

Conditional graph generation creates graphs with specified properties through property-conditioned models.

conditional independence, time series models

Conditional independence testing identifies which time series are independent given others in multivariate systems.

conditional position encoding in vit, computer vision

Position encoding dependent on content.

conditional position encoding, transformer

Position encoding conditioned on input.

conditioner,cmp

Diamond disk that maintains pad surface texture.

conditioning mechanisms, generative models

Ways to control generation.

conductive afm,metrology

Measure conductivity at nanoscale.

conductive anodic filament, caf, reliability

Dendrite causing shorts.

conductive flooring, manufacturing operations

Conductive flooring safely dissipates static charge from personnel.

conductive vs dissipative materials, facility

ESD protection material types.

conductive vs static-dissipative packaging, packaging

Different ESD protection levels.

conductivity measurement, manufacturing equipment

Conductivity measurement detects ionic contamination in ultrapure water.

conference,neurips,icml,paper

Top AI conferences: NeurIPS, ICML, ICLR, ACL. Papers show cutting edge. Arxiv for preprints.

confidence calibration,ai safety

Align predicted probabilities with actual accuracy.

confidence interval bootstrap, quality & reliability

Bootstrap confidence intervals use resampling to estimate parameter uncertainty.

confidence interval, quality & reliability

Confidence intervals provide ranges likely containing true parameter values.

confidence intervals for cpk, spc

Uncertainty in capability estimates.

confidence intervals for reliability, reliability

Statistical uncertainty in estimates.

confidence levels in reliability, reliability

Statistical confidence in estimates.

confidence penalty, machine learning

Penalize overconfident predictions.

confidence thresholding,ai safety

Only predict above confidence threshold.

confident learning,data quality

Identify and correct label errors.

confidential computing,privacy

Protect data in use with hardware-based isolation.

config,environment,secrets

Store config in env vars or config files, not code. Use secret managers (Vault, AWS Secrets) for API keys.

conflict minerals, environmental & sustainability

Conflict minerals regulations require supply chain due diligence ensuring materials don't fund armed conflict.

confocal microscopy,metrology

High-resolution 3D imaging.