TY - GEN
T1 - DOVETAIL - An abstraction for classical planning using a visual metaphor
AU - Magnaguagno, Mauricio Cecflio
AU - Pereira, Ramon Fraga
AU - Meneguzzi, Felipe
N1 - Publisher Copyright:
© 2016, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved.
PY - 2016
Y1 - 2016
N2 - While domain descriptions are often shared and manipulated through diagrams, most complex domains are still described using text-based languages. Code becomes an intermediary between the real-world and an abstract idea, and the programmer is merely a converter of diagrams into code. For automated planning this is no different. The state transition function is described in terms of a textual representation of actions and, although simple actions require little effort to define by the user, the learning process is often slow. New users have no metaphor to help them to visualize the domain description that they are working on and little information about why a planner fails due to formalization errors. In this paper, we propose a visual abstraction for both the planning domain actions and the planning process itself, to facilitate the design of classical planning domains. Using this abstraction, we expect to improve the learning curve for defining and subsequently diagnosing problems with new planning domains.
AB - While domain descriptions are often shared and manipulated through diagrams, most complex domains are still described using text-based languages. Code becomes an intermediary between the real-world and an abstract idea, and the programmer is merely a converter of diagrams into code. For automated planning this is no different. The state transition function is described in terms of a textual representation of actions and, although simple actions require little effort to define by the user, the learning process is often slow. New users have no metaphor to help them to visualize the domain description that they are working on and little information about why a planner fails due to formalization errors. In this paper, we propose a visual abstraction for both the planning domain actions and the planning process itself, to facilitate the design of classical planning domains. Using this abstraction, we expect to improve the learning curve for defining and subsequently diagnosing problems with new planning domains.
UR - http://www.scopus.com/inward/record.url?scp=85003828396&partnerID=8YFLogxK
M3 - Published conference contribution
AN - SCOPUS:85003828396
T3 - Proceedings of the 29th International Florida Artificial Intelligence Research Society Conference, FLAIRS 2016
SP - 74
EP - 79
BT - Proceedings of the 29th International Florida Artificial Intelligence Research Society Conference, FLAIRS 2016
A2 - Markov, Zdravko
A2 - Russell, Ingrid
PB - AAAI Press
T2 - 29th International Florida Artificial Intelligence Research Society Conference, FLAIRS 2016
Y2 - 16 May 2016 through 18 May 2016
ER -