← Back to AI Factory Chat

AI Factory Glossary

9,967 technical terms and definitions

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z All
Showing page 174 of 200 (9,967 entries)

synthetic data,data augmentation,gen

Synthetic data generated by models augments training. Helps with rare cases, privacy, and scale.

synthetic data,generation,privacy

Synthetic data generation preserves privacy. Train on fake data.

synthetic media detection,detection

Recognize AI-generated content.

synthetic patient generation, healthcare ai

Create realistic synthetic medical data.

synthetic reasoning data, data generation

Generated reasoning examples.

system audit, quality & reliability

System audits evaluate entire quality management system effectiveness.

system card,documentation

Comprehensive documentation of AI system.

system design,scale,high availability

I can walk you through system design interviews or real architectures: load balancers, caching, sharding, and HA/failover.

system message, prompting

High-level instructions for model behavior.

system message, prompting techniques

System messages set behavior guidelines and context for conversation.

system prompt extraction,ai safety

Attempt to recover hidden system instructions.

system prompt,instruction,setup

System prompt sets model behavior and context. First message in conversation. Defines persona.

system prompt,persona,role

System prompt defines the model behavior (role, style, boundaries). It is the always-on instruction that user prompts must follow.

system reliability modeling, reliability

Predict system-level reliability.

system scaling, business & strategy

System scaling improves performance through integration and architecture rather than shrinking.

system-in-package (sip),system-in-package,sip,advanced packaging

Integrate multiple dies or components in single package.

system-technology co-optimization, stco, design

Optimize across system architecture and technology.

systematic defect pattern, manufacturing operations

Systematic defect patterns repeat predictably indicating equipment or design issues.

systematic defect, yield enhancement

Systematic defects occur reproducibly at specific locations from design or process interactions.

systematic defects,metrology

Repeating defect patterns.

systematic sampling, quality & reliability

Systematic sampling selects every nth unit providing convenient coverage.

systematic signature, metrology

Repeating pattern across wafers.

systematic variation, design & verification

Systematic variations follow predictable patterns from layout-dependent effects.

systematic vth variation, device physics

Predictable threshold voltage shifts.

systematic within-die variation, manufacturing

Predictable WID patterns.

systematic yield loss, production

Predictable repeatable yield loss.

systematic yield loss, yield enhancement

Systematic yield loss results from deterministic process variations or design marginalities affecting specific die locations.

systolic array,matrix multiply

Systolic arrays efficiently compute matrix multiplications. Used in TPUs. Data flows through processing elements.

t-closeness, training techniques

T-closeness ensures sensitive attribute distribution within groups matches overall distribution.

t-closeness,privacy

Distribution of sensitive attributes similar across groups.

t-gate,rf design

Gate shape for high-frequency transistors.

t-squared chart, spc

Multivariate control chart.

t-test, quality & reliability

T-tests compare means between groups or against reference values.

t0, t0, foundation model

Zero-shot instruction-following model.

t0, t0, training techniques

T0 trains models on prompted datasets converted to unified text-to-text format.

t2i-adapter, generative models

Lightweight control adapter.

t2t-vit (tokens-to-token vit),t2t-vit,tokens-to-token vit,computer vision

Improved tokenization for ViT.

t5 (text-to-text transfer transformer),t5,text-to-text transfer transformer,foundation model

Treats all NLP tasks as text-to-text problems.

taaf, taaf, business & standards

Test Analyze and Fix methodology systematically eliminates failure modes improving reliability.

table-to-text,nlp

Describe tabular data in text.

tabnine,completion,local

Tabnine is AI code completion. Local model option for privacy.

tacotron, audio & speech

Tacotron is an end-to-end text-to-speech model that synthesizes speech from text using attention-based sequence-to-sequence learning.

tacotron2, audio & speech

Tacotron 2 improves on Tacotron by predicting mel-spectrograms with WaveNet vocoder for high-quality natural speech synthesis.

tagnn, tagnn, recommendation systems

Target Attentive Graph Neural Networks apply target-aware attention to session graphs for improved next-item recommendation.

taguchi methods,doe

Robust design using orthogonal arrays.

tail-free sampling, tfs, text generation

Remove low-probability tail.

take-back program, environmental & sustainability

Take-back programs enable manufacturers to collect used products for recycling or proper disposal.

takt time, manufacturing operations

Takt time is available production time divided by customer demand rate setting pace.

takt time, production

Rate of customer demand.

talking heads attention, transformer

Learn to redistribute attention across heads.