← 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 167 of 200 (9,967 entries)

split learning, training techniques

Split learning partitions models across parties computing collaboratively without sharing data.

split lot, manufacturing operations

Split lots divide wafers for experimental conditions or parallel processing paths.

split lot, production

Run part of lot with change for comparison.

split-cv, yield enhancement

Split capacitance-voltage separates electron and hole accumulation measuring interface properties.

split-cv,metrology

Capacitance-voltage measurement for interface states.

split-lot experiments, production

Compare process variations.

split-plot design, doe

Handle hard-to-change and easy-to-change factors.

splitfed learning, federated learning

Combine split and federated learning.

spm (sulfuric peroxide mixture),spm,sulfuric peroxide mixture,clean tech

Sulfuric acid + H2O2 for photoresist stripping and cleaning.

sporadic loss, manufacturing operations

Sporadic losses are sudden unusual events causing obvious performance degradation.

sports commentary generation,content creation

Create play-by-play commentary.

spos, spos, neural architecture search

Single Path One-Shot NAS trains supernet uniformly by sampling single paths improving architecture ranking accuracy.

spot instance management, infrastructure

Handle spot instance interruptions.

spot instance,preemptible,cost

Spot/preemptible instances are 60-90% cheaper but can be terminated. Good for fault-tolerant batch training.

spray cooling, thermal management

Spray cooling atomizes liquid into droplets that impinge and evaporate on surfaces for extreme heat removal.

spray processor, manufacturing equipment

Spray processors dispense chemicals as fine mist for uniform wafer treatment.

spreading resistance profiling, srp, metrology

Measure resistivity vs depth.

spreading resistance, thermal management

Thermal spreading resistance increases when heat spreads from small die to larger heat spreader.

sprint capacity, manufacturing operations

Sprint capacity temporarily increases output responding to demand spikes or recovering from disruptions.

spurious correlations, robustness

Models learn shortcuts instead of true patterns.

sputter yield,pvd

Number of target atoms ejected per incident ion.

sputtering,pvd

Bombard target with ions to eject atoms that deposit on wafer.

sqil, sqil, imitation learning

Combine IL and RL.

sqil, sqil, reinforcement learning advanced

Soft Q Imitation Learning simplifies imitation learning by labeling expert demonstrations with reward one and agent experience with reward zero.

sql generation,code ai

Generate database queries from natural language.

sql generation,natural language,query

Generate SQL from natural language. Text-to-SQL models.

sql generation,text to sql,nl2sql

Text-to-SQL converts natural language to SQL queries. Requires schema understanding. Fine-tune or few-shot for accuracy.

sql,database,postgres,timescaledb

I can help you design schemas, write SQL queries, optimize indexes, and explain Postgres/TimescaleDB features in simple terms.

squad, evaluation

Reading comprehension benchmark.

squad, evaluation

Stanford Question Answering Dataset tests reading comprehension.

square attack, ai safety

Query-efficient black-box attack.

squeeze excite,channel attention,se

Squeeze-and-Excite adds channel attention. Recalibrate features. Small overhead, accuracy gain.

squeeze-and-excitation, se, computer vision

Channel-wise attention mechanism.

squeeze-excitation, model optimization

Squeeze-and-excitation blocks recalibrate channel-wise features through global pooling and gating.

sr-gnn, sr-gnn, recommendation systems

Session-based Recommendation with Graph Neural Networks uses gated graph neural networks to model item transitions within sessions.

sraf (sub-resolution assist features),sraf,sub-resolution assist features,lithography

Tiny features on mask to improve main pattern.

sram yield, sram, manufacturing

Yield of memory arrays sensitive to variation.

srgnn variants, srgnn, recommendation systems

Session-based recommendation with graph neural networks includes multiple architectural variants for capturing transitions.

srnn, srnn, time series models

Stochastic Recurrent Neural Network uses stochastic hidden states for modeling uncertainty in sequences.

sse,server sent,streaming

Server-Sent Events stream data from server. One-way. Good for LLM token streaming.

ssim, ssim, evaluation

Structural similarity measure.

sso, sso, signal & power integrity

Simultaneous Switching Output noise occurs when multiple drivers switch creating transient currents that couple into power and signal nets.

ssta, ssta, design & verification

Statistical Static Timing Analysis propagates delay distributions through timing graphs.

ssvm multi-class, ssvm, structured prediction

Multi-class structured SVMs handle problems with structured inputs and multiple output classes using joint feature representations.

stability, metrology

Consistency over time.

stability, quality & reliability

Measurement stability is consistency of bias over time.

stability,metrology

Consistency of measurements over time.

stable diffusion architecture, generative models

Open-source latent diffusion model.

stable diffusion, multimodal ai

Stable Diffusion generates images through latent space diffusion with text conditioning.

stable diffusion,latent,text to image

Stable Diffusion generates images from text. Latent diffusion for efficiency. Open source, highly customizable.