AI Factory Glossary

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.

tally sheet, quality & reliability

Tally sheets record event counts or observations for frequency analysis.

tan barrier,beol

Tantalum nitride barrier.

tap controller, tap, advanced test & probe

Test Access Port controller is a state machine that interprets boundary scan instructions and controls test data register operations.

tape out,gdsii,foundry

Tape-out: final design sent to foundry as GDSII file. Point of no return. Months to get chips back.

tape width, packaging

Standard widths 8mm 12mm 16mm etc.

tapeout, business & strategy

Tapeout completes design sending data for mask fabrication.

tapeout,design

Finalize design and send to foundry for mask making.

tarc (top arc),tarc,top arc,lithography

ARC layer on top of resist.

target encoding,mean,category

Target encoding uses target mean per category. Can leak information.

target impedance, signal & power integrity

Target impedance specifies maximum allowable power distribution network impedance versus frequency ensuring acceptable voltage ripple.

target speaker extraction, audio & speech

Target speaker extraction isolates specific speaker from mixture using enrollment utterance.

target thickness, process

Desired final wafer thickness.

target tracking, manufacturing operations

Target tracking adjusts process aims as equipment drifts maintaining centered output.

target value engineering, quality

Design to hit nominal not just meet specs.

target value, spc

Desired nominal value.

target,pvd

Source material being sputtered (metal alloy).

task allocation, ai agents

Task allocation assigns responsibilities to agents based on capabilities and load.

task and motion planning (tamp),task and motion planning,tamp,robotics

Combine high-level planning with motion.

task arithmetic, model merging

Add/subtract task vectors.

task balancing, multi-task learning

Ensure fair task representation.

task decomposition, ai agents

Task decomposition breaks complex goals into manageable subtasks.

task diversity, training techniques

Task diversity in instruction tuning exposes models to varied problem types.

task grouping, multi-task learning

Cluster related tasks.

task instruction, prompting techniques

Task instructions specify desired actions clearly guiding model behavior.

task interference, multi-task learning

Tasks hurting each other's performance.

task prompting, multi-task learning

Use prompts to specify task.

task recognition in icl, theory

Model identifying task from examples.

task routing, multi-task learning

Direct inputs to task-specific modules.