Amortized Inference
An efficient paradigm for variational inference that trains a shared inference network to map observations to variational parameters, reducing parameter count from O(n) to a fixed network size while enabling one-step generalization to new observations.
Timeline (last 90 days)
This framework amortizes all variational parameters through a shared neural conditioner.
Amortized inference originates from the design paradigm of variational autoencoders (VAEs), using a shared inference network to reduce the parameter count from O(n) to the network's fixed parameter count, while enabling one-step inference generalization to new observation points.
All Facts (2)
This framework amortizes all variational parameters through a shared neural conditioner.
50%UnverifiedAmortized inference originates from the design paradigm of variational autoencoders (VAEs), using a shared inference network to reduce the parameter count from O(n) to the network's fixed parameter count, while enabling one-step inference generalization to new observation points.
50%