← Back to AI Factory Chat

AI Factory Glossary

1,106 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 5 of 23 (1,106 entries)

selective epitaxy, process integration

Selective epitaxy grows crystalline semiconductor only on exposed silicon surfaces not on dielectrics.

selective epitaxy,cvd

Epi grows only on exposed silicon not on oxide/nitride.

selective etch,metrology

Preferentially remove one material.

selective kernel networks, computer vision

Dynamically adjust receptive field.

selective knowledge distillation, model compression

Choose which knowledge to transfer.

selective prediction, ai safety

Selective prediction abstains from answering when confidence is insufficient.

selective prediction,ai safety

Choose when to abstain from prediction.

selective soldering, packaging

Solder specific areas.

selective ssm, llm architecture

Selective SSMs dynamically adjust state transitions based on input content.

selectivity (cmp),selectivity,cmp

Ratio of polish rates between different materials.

selectivity, metrology

Ability to measure target in presence of interferences.

selectivity,etch

Ratio of etch rates between target material and underlying layer.

self consistency,majority vote

Self-consistency samples multiple reasoning paths, takes majority vote answer. Improves accuracy over single sample.

self distillation,consistency,regularize

Self-distillation distills model to itself. Different views or augmentations. Regularization effect.

self host,on prem,local deploy

Self-hosting gives control over data and cost at scale. Requires ML ops expertise. Use for sensitive data.

self-aligned contact process, sac, process integration

Contacts aligned to spacers and gates.

self-aligned contact, process integration

Self-aligned contacts use spacers and selective etch to automatically align contact openings to gates.

self-aligned via (sav),self-aligned via,sav,beol

Via aligning itself to metal lines.

self-alignment, training techniques

Self-alignment improves models using their own generated critiques and revisions.

self-ask,reasoning

Model asks and answers sub-questions.

self-attention asr, audio & speech

Self-attention ASR uses transformer architecture for speech recognition without recurrent connections.

self-attention in capsules,neural architecture

Combine capsules with attention.

self-attentive hawkes, time series models

Self-attentive Hawkes processes combine transformer architectures with temporal point process likelihoods for scalable event sequence modeling.

self-consistency, prompting

Sample multiple reasoning paths and vote.

self-consistency, prompting techniques

Self-consistency samples multiple reasoning paths selecting most consistent final answer.

self-consistency,reasoning

Generate multiple reasoning paths and pick the most common answer.

self-critique, prompting

Model identifies flaws in its output.

self-critique, prompting techniques

Self-critique prompts models to identify weaknesses in generated responses.

self-critiquing, ai safety

Model critiques its own outputs.

self-distillation, model compression

Distill model into itself.

self-distillation, model optimization

Self-distillation uses model's own predictions as soft targets improving generalization.

self-distillation, self-supervised learning

Distill model into itself.

self-ensembling for domain adaptation, domain adaptation

Ensemble predictions for consistency.

self-evaluation,evaluation

Model evaluates its own outputs.

self-explaining neural networks, senn, explainable ai

Architectures designed for interpretability.

self-gating, neural architecture

Use gates controlled by input.

self-heating in soi,reliability

Temperature rise due to buried oxide.

self-heating interconnect, signal & power integrity

Self-heating in interconnects couples electrical and thermal effects influencing reliability.

self-heating modeling, simulation

Model temperature rise in device.

self-inspection, quality & reliability

Self-inspection has operators check their own work immediately.

self-instruct, data generation

Generate instructions from model.

self-instruct, training techniques

Self-Instruct generates training data by prompting models to create instructions and responses.

self-monitoring, ai agents

Self-monitoring tracks progress evaluating whether actions advance toward goals.

self-paced learning, advanced training

Self-paced learning automatically selects training samples based on current model capability starting with easier examples.

self-paced learning, machine learning

Let model select training order.

self-rag, rag

Model decides when and what to retrieve.

self-rag, rag

Self-RAG adaptively retrieves only when needed using learned retrieval decisions.

self-rag,rag

Model decides when it needs to retrieve documents vs generate from memory.

self-refine, prompting

Iteratively improve generations.

self-refine, prompting techniques

Self-refine iteratively generates critiques and improvements without external feedback.