Paper Outline

This note is a short scaffold for the main paper arc.

Related notes:

Problem framing
Contributions
CVAE objective and loss terms
Formalizing strategy identification

Proposed arc

Problem statement: factorization

p(y \mid x) \to (r(z \mid x), g(y \mid x, z)).

Methodology: CVAE variational objective and the factorization / disentanglement training loop.
Experiments.
Theory.

Theory subsection sketch

Formalization of the problem.
Relationship to established problems in prior work, such as identifying or learning finite mixture models.
Theoretical results in our setting and their connection to the proposed methodology.

Ordering rationale

Theory can appear later in the paper, after the problem, methodology, and experiments. That ordering makes the paper more accessible before moving into the more technical analysis.