Advances in Artificial Intelligence: 19th Conference of the by Luc Lamontagne, Mario Marchand

By Luc Lamontagne, Mario Marchand

This e-book constitutes the refereed complaints of the nineteenth convention of

the Canadian Society for  Computational reports of Intelligence, Canadian

AI 2006, held in Québec urban, Québec, Canada in June 2006.

The forty seven revised complete papers awarded have been conscientiously reviewed and chosen

from 220 submissions. The papers are equipped in topical sections on

agents, bioinformatics, constraint delight and dispensed seek,

knowledge illustration and reasoning, ordinary language, reinforcement

learning and, supervised and unsupervised learning.

Show description

Read Online or Download Advances in Artificial Intelligence: 19th Conference of the Canadian Society for Computational Studies of Intelligence, Canadian AI 2006, Quebec City, PDF

Similar computational mathematicsematics books

Groundwater Hydrology Conceptual and Computational Models

Groundwater is an important resource of water through the global. because the variety of groundwater investigations raise, it is very important know how to enhance accomplished quantified conceptual types and savour the root of analytical strategies or numerical equipment of modelling groundwater circulate.

Additional resources for Advances in Artificial Intelligence: 19th Conference of the Canadian Society for Computational Studies of Intelligence, Canadian AI 2006, Quebec City,

Example text

Fernie G. : A DecisionTheoretic Approach to Task Assistance for Persons with Dementia. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI’05), pages 12931299, Edinburgh, Scotland, 2005. 7. : Techniques for Plan Recognition, User Modeling and User AdaptedInteraction, (2001), (11) 31-48. 8. : A cognitive system for a smart home dedicated to people in loss of autonomy, In: Proc. of the 3rd International Conference on Smart homes and health Telematic, ICOST’05, Sherbrooke, Canada, (2005), 245-254.

For instance, considering a well-known Kautz’s example given in [10], we see that if we observe two actions GetGun and GotoBank, then we cannot automatically conclude that the observed agent A Smart Home Agent for Plan Recognition 27 wants to rob a bank, or deduce a disjunction of possible plans Hunt or RobBank as proposed by Kautz’s theory. The fact is that his intentions can be to go on a hunting trip and to cash a check on the way, knowing that the initial set of possible plans were RobBank and Hunt.

Bn+1 ◦ DisU (an , bn ) ◦ . . ◦ DisU (ak+1 , bk+1 ) ◦ ok ◦ . . ◦ o1 , if n ≤ m bn ◦ . . ◦ bm+1 ◦ DisU (am , bm ) ◦ . . ◦ DisU (ak+1 , bk+1 ) ◦ ok ◦ . . ◦ o1 , if m ≤ n where DisU is a disunification operation defined as an injective application: A ∪ V × A → A ∪ V , on the set of incomparable actions of plans α, β: DisU (a, b) = c if f ∃c ∈ A : c ≺p a and c ≺p b x elsewhere, with Sub(x) = {a ◦ b, b ◦ a} To summarize, the recognition process consists in finding a recognition space which is a minimal model of interpretation of the observations O that admits a supremum ∇sup , corresponding to the most specific common subsumer of all possible plans, and that admits an infinimum Δinf , corresponding to the minimal intention schema predicting the future behaviour of the observed agent.

Download PDF sample

Rated 4.01 of 5 – based on 16 votes