1 paper
Sohini Roychowdhury, Adam Holeman, Mohammad Amin +3
For online ad-recommendation systems, processing complete user-ad-engagement histories is both computationally intensive and noise-prone. We introduce Dynamix, a scalable, personal…