AI Factory Glossary

balance,wellbeing,sustainable

Sustainable pace over crunch. AI is marathon not sprint. Balance work with wellbeing.

balanced sampling, machine learning

Sample to balance class distribution.

ball shear, failure analysis advanced

Ball shear testing measures solder ball attach strength in BGA packages by shearing balls from pads.

barren plateaus, quantum ai

Training difficulty in quantum circuits.

bart (bidirectional and auto-regressive transformer),bart,bidirectional and auto-regressive transformer,foundation model

Combines bidirectional encoder with autoregressive decoder for seq2seq.

base model, llm architecture

Base models are foundational pre-trained models before any fine-tuning.

base model,instruct,chat

Base model is pretrained only. Instruct/chat versions are fine-tuned for dialog. Different use cases.

batch learning,machine learning

Train on fixed dataset offline.

batch size,model training

Number of examples processed together in one forward/backward pass.

batch wait time, operations

Time waiting for full batch.

batching inference, llm optimization

Batching combines multiple requests processing them together for efficiency.

bayesian change point, time series models

Bayesian change point detection models segment boundaries as latent variables with posterior inference.

bayesian neural networks,machine learning

Neural networks with probabilistic weights.

bayesian optimization,model training

Use probabilistic model to guide hyperparameter search efficiently.

bed-of-nails, failure analysis advanced

Bed-of-nails fixtures use arrays of spring-loaded pins to contact PCB test points for in-circuit testing and fault isolation.

beginner,explain like 5,simple

When you say explain like I am 5, I will use tiny steps, concrete examples, and avoid jargon, then we can gradually increase depth and formality.

behavioral testing, explainable ai

Test model behavior on synthetic data.

beit (bert pre-training of image transformers),beit,bert pre-training of image transformers,computer vision

Masked image modeling for ViT.

beit pre-training, computer vision

Predict discrete visual tokens.

bert (bidirectional encoder representations),bert,bidirectional encoder representations,foundation model

Masked language model for understanding tasks.

beta-vae,generative models

VAE with adjustable disentanglement.

bga x-ray, bga, failure analysis advanced

BGA X-ray inspection detects solder joint defects like voids bridges and insufficient wetting in ball grid array packages.

bias amplification, fairness

Models exaggerating biases in data.

bias mitigation strategies, fairness

Reduce bias in models.

bias mitigation,ai safety

Techniques to reduce unfair biases in model training data or outputs.

bidirectional attention,transformer

Each token can attend to all other tokens (BERT-style).

bidirectional language modeling, foundation model

Predict from both directions.

bigbird attention, llm architecture

Sparse attention with random global and local.

bigbird attention, llm optimization

BigBird uses sparse attention with local sliding window random and global connections.

bigbird,foundation model

Sparse attention with local global and random connections.

bignas, neural architecture search

BigNAS scales once-for-all training to extremely large search spaces using improved sampling strategies.

binarized neural networks (bnn),binarized neural networks,bnn,model optimization

Extreme quantization to 1-bit.

binary networks, model optimization

Binary networks constrain weights and activations to binary values enabling extreme compression.

binary neural networks,model optimization

Weights and activations are binary.

binding affinity prediction, healthcare ai

Predict strength of molecular binding.

biofilter, environmental & sustainability

Biofilters use microorganisms to biologically degrade VOCs and odorous compounds.

biomedical text mining,healthcare ai

Extract knowledge from medical literature.

bit diffusion, generative models

Diffusion at bit level for discrete data.

blackboard system, ai agents

Blackboard systems use shared knowledge space where agents post and read information.

blip (bootstrapping language-image pre-training),blip,bootstrapping language-image pre-training,multimodal ai

Vision-language understanding and generation.

blip-2,multimodal ai

Efficient vision-language pre-training.

block-recurrent transformer,llm architecture

Process input in blocks with recurrence between blocks.

block-wise model merging, model merging

Merge layer by layer.

blockqnn, neural architecture search

BlockQNN extends Q-learning NAS by searching architectures block-by-block reducing search space complexity.

bloom,foundation model

Multilingual open-source LLM.

board-level reliability, failure analysis advanced

Board-level reliability testing evaluates assembled package performance under thermal mechanical and environmental stress.

bohb, bohb, neural architecture search

Bayesian Optimization and HyperBand combines multi-fidelity optimization with Bayesian modeling for efficient NAS.

bom, bom, supply chain & logistics

Bill of Materials lists all components sub-assemblies and quantities required to manufacture a product.

boosting,machine learning

Sequentially train models on hard examples.

born-again networks, model compression

Repeatedly distill knowledge.