Paper revisions needed: State of the Field, Software Design, and References

[mbsuraj]

This issue is for JOSS Review #10653

For the JOSS sofware paper criteria, following paper sections need minor expansion:

1. State of the Field: name specific packages and add build-vs-contribute justificationThe current section refers to "several libraries" and "existing probabilistic programming frameworks" without naming them. JOSS requires that "when related tools exist, authors must provide a clear 'build vs. contribute' justification."

Please -
a. Name the specific existing tools in this space (e.g., NumPyro, Pyro, TensorFlow Probability, Laplace-Redux, or similar BDL packages)
b. Explain briefly why a standalone package was built rather than contributing a MILE implementation to an existing framework like BlackJAX or NumPyro

2. Software Design: expand with architectural trade-offs
   The current section is a single sentence about JAX parallelization. JOSS asks for "trade-offs weighed, the design/architecture chosen, and why it matters."

Please consider briefly addressing:

1. Why fully-connected networks only ?
2. Why full-batch training ?
3. The two-stage design (optimization → sampling) ?

I believe two to three additional sentences would satisfy this requirement.

[EmanuelSommer]

Thank you for the good feedback. We revised the paper to name representative related tools more explicitly, including posteriors/Laplace-Redux for optimization-based BDL workflows and NumPyro/BlackJAX for lower-level MCMC and probabilistic inference.

We also added a short Software Design clarification as well.

[mbsuraj]

Thank you for taking time and adding more context to the paper. This meets the criteria I was tracking to for the paper.