classifier-free guidance does: interpolates between conditional and unconditional generation

Boosting (machine learning)

Boosting reduces bias in ML models

branch prediction does: guesses which way an if-statement will go to keep the pipeline full

Branch predictors guess the outcome of conditional jumps to keep the pipeline full

soft targets carry more information than hard labels: they encode class similarities

Why do some learning methods need to explore more than others?

instruction tuning does: fine-tunes on (instruction, response) pairs

Fine-tuning adapts pre-trained models to new tasks

Adam vs SGD: Adam adapts per-parameter rates, SGD often generalizes better with tuning

Adam adjusts learning rates per-parameter, SGD generalizes better with tuning

to standardize: when you need zero mean and unit variance for gradient-based optimization

Why do we need to make data uniform before training a model?