The QuasiHyperbolic Momentum Algorithm (QHM) is a simple alteration of momentum SGD, averaging a plain SGD step with a momentum step. QHAdam is a QH augmented version of Adam, where we replace both of Adam's moment estimators with quasihyperbolic terms. QHAdam decouples the momentum term from the current gradient when updating the weights, and decouples the mean squared gradients term from the current squared gradient when updating the weights.
In essence, it is a weighted average of the momentum and plain SGD, weighting the current gradient with an immediate discount factor $v_{1}$ divided by a weighted average of the mean squared gradients and the current squared gradient, weighting the current squared gradient with an immediate discount factor $v_{2}$.
$$ \theta_{t+1, i} = \theta_{t, i}  \eta\left[\frac{\left(1v_{1}\right)\cdot{g_{t}} + v_{1}\cdot\hat{m}_{t}}{\sqrt{\left(1v_{2}\right)g^{2}_{t} + v_{2}\cdot{\hat{v}_{t}}} + \epsilon}\right], \forall{t} $$
It is recommended to set $v_{2} = 1$ and $\beta_{2}$ same as in Adam.
Source: Quasihyperbolic momentum and Adam for deep learningPaper  Code  Results  Date  Stars 

Component  Type 


🤖 No Components Found  You can add them if they exist; e.g. Mask RCNN uses RoIAlign 