Information Technology Reference
In-Depth Information
The smart games stimulates the comprehension in terms of: the story characters
(who games), the events that take place in the story (what games), the temporal
ordering of such events (before/after, while/after, before/while, before/while/after
games), and the causal relations (cause, effect, cause/effect games) [4, 5]. The
adopted game framework is a generalisation of the EMAPPS one [6].
Fig. 1 The smart games generation workflow
Fig. 1 summarises the smart games generation process. The process is made up
of the four phases described below:
Phase A. Stories are stored as flat text in a story repository. From a story text,
an NLP module generates an annotated story. The annotations follow a variant
of the TimeML language [7], that was extended in [8] so to include new tags
useful for the TERENCE smart games. For instance, the ENTITY and CLINK
tags aim, respectively, at representing the entity related to an event (the EVENT
tag is already present in the TimeML standard), and the causal relation between
two events. After the annotation process, the annotated story is stored in the same
repository;
Phase B. The output of the annotation module is initially checked by a reasoner
in terms of consistency of the annotations, i.e., by detecting any eventual tem-
poral inconsistency (e.g., if a relation between two events E 1 and E 2 is that E 1
is before E 2 , there cannot exist other relations stating that E 1 is after E 2 ). If
inconsistencies are found, the annotations cannot be stored 1 .Otherwise,thean-
notations are enriched by adding deduced temporal relations as further TLINK
tags [9]. This new consistent and enriched story is also stored in the story repos-
itory;
Phase C. Starting from the consistent and enriched story, the reasoner module
generates automatically instances of smart games. For instance, for a WHO-game
(i.e., who is doing something in the story) related to a certain event, hereafter
called fixed event :
1
Techniques of relaxation are under development, that will lead to a “self-reparation” process
concerning the inconsistent annotations.
 
Search WWH ::




Custom Search