#ag
> Is Bayesian a belif?
Bayesian is like the model (not a belief).
> Bayesian de-convolution?
Fig.2.10. (BDA) is similar to deconvolution happending in f = g * N(0,1). Convolution is related to kernel, fourier transform etc.
> Opinion on reconstruction conjecture?
We choose high dimensional model with enough structure so that it can be identified. Similarly, random is a deterministic in high dimension projected to lower dimension due to finiteness of our perception -- finite implemeted program.
> What is the difference between verification vs identification?
Similar but the verification is more mathematical and well-defined.
> Meaning of “Mathematical science is an indivisible whole, an organism whose vitality is conditioned upon the connection of its parts” by Hilbert?
Hierarchical modeling can come before Bayesian.
e.g. Euclidean 5 axioms connected to each other implying the othere.g. Euclidean 5 axioms connected to each other implying the other
- Neural network is not preserved under linear transformation due to regularization which is isn't preserved under prior (rotation is also involved).
![[AngieAxiom.pdf]]