AI Factory Glossary

solder paste printing, manufacturing

Apply paste to PCB.

solder paste volume, manufacturing

Amount deposited per pad.

solder reflow, packaging

Melt solder to form connection.

solder tim, thermal management

Solder thermal interface materials provide permanent low-resistance thermal connections in high-performance packages.

solder voids x-ray, quality

Gas pockets in solder.

solenoid valve, manufacturing equipment

Solenoid valves provide fast on-off control for chemical delivery.

solubility prediction, chemistry ai

Predict compound solubility.

solvent distillation, environmental & sustainability

Solvent distillation separates and purifies organic solvents based on boiling point differences.

solvent exhaust,facility

Exhaust system for organic solvents and VOCs.

solvent recovery, environmental & sustainability

Solvent recovery systems capture and purify organic solvents from process exhaust enabling reuse and reducing disposal costs.

solvent recovery,facility

Recycle and reuse solvents from cleaning processes.

song lyrics generation,content creation

Write lyrics for music.

sophia, optimization

Second-order optimizer for language models.

sort pooling, graph neural networks

Sort pooling orders nodes by learned structural roles enabling use of 1D convolutions for graph classification.

sort yield, production

Yield at wafer sort.

sort, manufacturing operations

Sort operations separate wafers based on test results or characteristics.

sort, manufacturing operations

Sort removes unnecessary items from workplace retaining only what is needed.

sorter,automation

Machine that sorts and organizes wafers by lot or specification.

sortpool variant, graph neural networks

SortPool variants improve graph classification by learning node importance for sorted feature concatenation.

sound source localization, multimodal ai

Locate sound sources using vision.

soundstream, audio & speech

SoundStream is a neural audio codec using residual vector quantization for high-quality compression.

source filtering, rag

Filter by document source.

source inspection, quality & reliability

Source inspection checks conditions causing defects before they occur.

source latency, design & verification

Source latency is delay from original clock generation to tree root.

source power,etch

RF power for generating plasma density.

source separation, audio & speech

Audio source separation decomposes mixed signals into constituent sources like speech instruments or noise.

source separation,audio

Separate mixed audio into individual sources (vocals instruments).

source-free domain adaptation, domain adaptation

Adapt without access to source data.

sourcegraph cody,code,context

Cody is Sourcegraph AI with codebase context. Understands your code.

space charge region, device physics

Region depleted of mobile carriers.

space-filling designs, doe

Cover parameter space evenly.

spacer formation, process integration

Spacer formation deposits and etches insulating material beside gate structures controlling overlap and junction profiles during source-drain implantation.

span boundary objective, nlp

Predict span boundaries.

span masking, nlp

Mask contiguous spans of tokens.

span-based parsing, structured prediction

Span-based parsing predicts constituency trees by scoring all possible spans and their labels using neural representations.

spare parts inventory,operations

Stock of replacement components.

spare parts, manufacturing operations

Spare parts inventory supports rapid equipment repair reducing downtime.

spark,big data,distributed

Apache Spark processes big data distributed. PySpark for Python. ML pipelines at scale.

sparse attention,transformer

Attention pattern where each token only attends to a subset of tokens rather than all.

sparse autoencoder,feature,decompose

Sparse autoencoders decompose activations into interpretable features. Scale to large models.

sparse autoencoder,sae,features

Sparse autoencoders find interpretable features in model activations. Tool for understanding what models learn.

sparse autoencoders for interpretability, explainable ai

Extract interpretable features.

sparse mapping, robotics

Use sparse feature points.

sparse mixture, llm architecture

Sparse mixture of experts activates subset of experts per input.

sparse model,model architecture

Model where only subset of parameters activate (MoE).

sparse retrieval, rag

Use keywords (BM25 TF-IDF).

sparse retrieval, rag

Sparse retrieval uses term matching like BM25 for document retrieval.

sparse training, model optimization

Sparse training maintains sparsity throughout training never materializing dense networks.

sparse transformer patterns, transformer

Different sparsity patterns for attention.

sparse upcycling,model architecture

Convert dense model to MoE by splitting weights into experts.