Back to timeline

Research · December 2014

The Adam optimizer

Kingma and Ba described a method that picks a learning rate for each parameter separately from estimates of the first and second moments of the gradient.

Why it matters

Tuning training stopped being a craft: Adam worked acceptably almost everywhere without a parameter search.

Before Adam the learning rate was a leading reason a network failed to train, and choosing it took experience. Adam adapts the step for each weight and corrects the bias in its estimates early in training. It is not the best method in every case but it is a reliable default. It is among the most cited papers in machine learning precisely for being practical.

Event record

Event date
December 2014
Timeline date
Event date
Verification
Sources gathered automatically · September 17, 2026
Lines
ID
evt-0226

Preprint of 22 December 2014; presented at ICLR 2015.

Sources

Related events