Coupling conceptual modeling and rules for the annotation of dramatic media

Tracking #: 680-1890

Vincenzo Lombardo1

Responsible editor: 
Guest Editors RR2013 Special Issue

Submission type: 
Full Paper
This paper presents an ontological approach to the domain of drama. After a description of the drama domain in a cross-cultural and media setting, we introduce the ontology Drammar. Drammar consists of two com- ponents, encoding the conceptual model and SWRL rules, respectively. The conceptual model, mainly grounding on AI theories, represents the major concepts of drama, such as agents, actions, plans, units, emotions and values. Then, the paper focuses on the rule component, that augments the representation by mapping the intentions of the characters onto the actions actually performed and by appraising the emotion felt by the characters in the drama. To illustrate the functioning of the ontology we introducing a running example from an excerpt of the drama Hamlet. Finally, we carry out an evaluation of the approach on an annotation task that is relevant for drama studies research and teaching. In particular, the emotion appraisal is tested on the main characters of four dramas of di erent nature, with precision and recall results with respect to a human annotator.
Full PDF Version: 

Minor Revision

Solicited Reviews:
Click to Expand/Collapse
Review #1
Anonymous submitted on 13/Jun/2014
Review Comment:

Authors have taken into account and addressed all issues raised in the previous review. The main issue was evaluation and authors added the requested section. Only a typo (missing reference in page 2, right, 2nd paragraph) must be fixed.

Review #2
By Alessandro Oltramari submitted on 30/Jun/2014
Review Comment:

The authors adequately addressed all the issues I highlighted in my review. Overall, I think that the paper has now reached a good level of maturity and is suitable for publication.

Review #3
By Pablo Gervás submitted on 05/Jul/2014
Minor Revision
Review Comment:

The paper has improved considerably as a result of the last revision. The description of the rules is now more consistent (though still significantly difficult to understand).

Having the annotation workflow explicitly in the paper helps the reader to place conceptually the various pieces within the overall process.

I am willing to accept the statement by the authors that the ontology as presented is good for the particular task for which it was created, and that it might need extensions to make it suitable for other tasks beyond that one.

The effort made to separate material borrowed from AI improves the paper.

The section on SWRL rules still suffers somewhat from interleaving of review of tecnologies and authors' contribution but I guess it is difficult to solve that without incurring in additional overall complexity for the whole paper.

The explanation of the rules for assigning emotions to characters has become frighteningly long and complex, but it has definitely improved with respect to the previous effort. The new tree representation is a little more intuitive to follow, though still somewhat obscure for the uninitiated.

The systematic evaluation is a very welcome addition.

Overall, I think the paper is now much better than it was. I also consider that the authors have made a comendable effort to address the comments given by the reviewers. In consequence, I consider the paper acceptable for publication. Nonetheless, there are some small details that may require minor revision. A short list follows:

- page 2, column 2, 2nd paragraph: missing internal reference appears as ??
- page 3, column 2, end of first paragraph: "perspective because links the drama" -> "perspective because they link the drama"?
- page 7, section heading: appears as "Drama Ontology"; should it be "Drammar"? The authors talk both about a "Drama Ontology" and the "Drammar Ontology". Are these different? The same? Is one a description and the other a name? Some careful rewriting may avoid the ambiguity.
- page 7, column 2, end of first paragraph: "a consolidate agent model" -> "a consolidated agent model"
- page 14, column 2, "different than the unit ordering" -> "different from the unit ordering"
- page 19, column 2 paragraph 2: "Love and Hate emotion types are still annotated manually" Why? The authors should explain whether this is a limitation of their procedure, a peculiarity of these emotion types. I understand that these emotion types have then been excluded from the evaluation described in section 6. If so, this should be mentioned explicitly. This point should also be discussed as a possibly important shortcoming of the process. Surely Love and Hate constitute crucial emotional ingredients of every drama. Please address this comment in the final version of the paper.
- page 20, column 1, towards end of paragraph 1: "who is that feels" -> "who it is that feels"
- page 20, column 1, paragraph 4: "human annotators tend to be more elliptical in the annotation, rather than addressing actual differences" This observation was cryptic to me, and I only managed to understand it after reading through to the rest of the section. After doing that, I managed to understand what is meant by "tend to be more elliptical" (though I think a clearer formulation of that idea would surely be possible and desireable). But I can still make no sense of the "rather than addressing actual differences" part. What differences? How would human annotators address them? Where? When? Please clarify.