CS2125 Paper Review Form - Winter 2018

Reviewer: Mikhail Berezovskiy

Paper Title: The next evolution of MDE: a seamless integration of machine learning into domain modeling

Author(s): Thomas Hartmann · Assaad Moawad · Francois Fouquet · Yves Le Traon

1) Is the paper technically correct?
 [v] Yes
 [ ] Mostly (minor flaws, but mostly solid)
 [ ] No

2) Originality
 [ ] Very good (very novel, trailblazing work)
 [ ] Good 
 [v] Marginal (very incremental)
 [ ] Poor (little or nothing that is new)

3) Technical Depth
 [ ] Very good (comparable to best conference papers)
 [ ] Good (comparable to typical conference papers)
 [v] Marginal depth
 [ ] Little or no depth

4) Impact/Significance
 [ ] Very significant
 [ ] Significant
 [v] Marginal significance.
 [ ] Little or no significance.

5) Presentation
 [ ] Very well written
 [ ] Generally well written
 [v] Readable
 [ ] Needs considerable work
 [ ] Unacceptably bad

6) Overall Rating
 [ ] Strong accept (award quality)
 [ ] Accept (high quality - would argue for acceptance)
 [v] Weak Accept (borderline, but lean towards acceptance)
 [ ] Weak Reject (not sure why this paper was published)


7) Summary of the paper's main contribution and rationale for your recommendation. (1-2 paragraphs)
	The concept of integration of machine learning into a domain modeling following a current trend of 
achieving more robust and sustainable systems. Motivation is clear. In particular application modeling language is 
focusing solving a particular problem with Smart Grids and quite narrowly focused on the application.
	The article quite overloaded with formalism, and struggle with examples, it's not clear how learned 
and derived features actually works in the system. As a result of the paper, authors compare ML algorithms, 
but not the actual integration of ML into models and how models are changed in the process.

8) List 1-3 strengths of the paper.  (1-2 sentences each,
identified as S1, S2, S3.)
 - There is a complete working system, with ongoing developments, and commercialized
 - Despite the quite narrow application of the system, it's still showed valid results

9) List 1-3 weaknesses of the paper (1-2 sentences each,
identified as W1, W2, W3.)
- I think integration with external ML platforms and libraries through models will make it more seamless 
than the selection of predefined algorithms
- Lack of examples in the paper
- Absent or not showed an integration of models with the running environment