34 resultados para Translation strategies
Resumo:
We generalise and extend the work of Iñarra and Laruelle (2011) by studying two person symmetric evolutionary games with two strategies, a heterogenous population with two possible types of individuals and incomplete information. Comparing such games with their classic homogeneous version vith complete information found in the literature, we show that for the class of anti-coordination games the only evolutionarily stable strategy vanishes. Instead, we find infinite neutrally stable strategies. We also model the evolutionary process using two different replicator dynamics setups, each with a different inheritance rule, and we show that both lead to the same results with respect to stability.
Resumo:
8 p.
Resumo:
Recent works in the area of adaptive education systems point out the importance of aumenting the student model to improve the personalization and adaptation to the learner by means of several aspects such as emotions, user locations or interactions. Until now the study of interactions has been mainly focused on the student-learning system flow, despite the fact that the most successful and used way of teaching are the traditional face-to-face interactions. In this project, we explore the use of interactions among teachers and students, as they occur in traditional education, to enrich the current student models, with the aim of providing them with useful information about new characteristics for improving the learning process. At a first step, in this paper we present the formal process carried out to obtain information about teachers’ expertise and necessities regarding the direct interactions with students.
Resumo:
Methods for generating a new population are a fundamental component of estimation of distribution algorithms (EDAs). They serve to transfer the information contained in the probabilistic model to the new generated population. In EDAs based on Markov networks, methods for generating new populations usually discard information contained in the model to gain in efficiency. Other methods like Gibbs sampling use information about all interactions in the model but are computationally very costly. In this paper we propose new methods for generating new solutions in EDAs based on Markov networks. We introduce approaches based on inference methods for computing the most probable configurations and model-based template recombination. We show that the application of different variants of inference methods can increase the EDAs’ convergence rate and reduce the number of function evaluations needed to find the optimum of binary and non-binary discrete functions.
Resumo:
11 p.
Resumo:
Eguíluz, Federico; Merino, Raquel; Olsen, Vickie; Pajares, Eterio; Santamaría, José Miguel (eds.)
Resumo:
Eguíluz, Federico; Merino, Raquel; Olsen, Vickie; Pajares, Eterio; Santamaría, José Miguel (eds.)
Resumo:
Eguíluz, Federico; Merino, Raquel; Olsen, Vickie; Pajares, Eterio; Santamaría, José Miguel (eds.)
Resumo:
Eguíluz, Federico; Merino, Raquel; Olsen, Vickie; Pajares, Eterio; Santamaría, José Miguel (eds.)
Resumo:
Eguíluz, Federico; Merino, Raquel; Olsen, Vickie; Pajares, Eterio; Santamaría, José Miguel (eds.)
Resumo:
Eguíluz, Federico; Merino, Raquel; Olsen, Vickie; Pajares, Eterio; Santamaría, José Miguel (eds.)
Resumo:
Santamaría, José Miguel; Pajares, Eterio; Olsen, Vickie; Merino, Raquel; Eguíluz, Federico (eds.)
Resumo:
Santamaría, José Miguel; Pajares, Eterio; Olsen, Vickie; Merino, Raquel; Eguíluz, Federico (eds.)
Resumo:
Santamaría, José Miguel; Pajares, Eterio; Olsen, Vickie; Merino, Raquel; Eguíluz, Federico (eds.)
Resumo:
Santamaría, José Miguel; Pajares, Eterio; Olsen, Vickie; Merino, Raquel; Eguíluz, Federico (eds.)