Ph.D. student in computer sciences, Universidade de São Paulo
https://www.ime.usp.br/~eguetti/
The Semantic Web represents the greatest evolution in information connectivity. Enabling the connection of different sources of resources and allowing computers to perform sophisticated tasks for our interest. Our goal is to introduce the Semantic Web through its evolution in the World Wide Web in knowledge management, and particularly in the context of large corporations. Addressing the main topics of study within semantic technology, as well as the challenges in the development of semantic applications, and finally presenting case studies of success.
MSc(c) - Universidad Católica San Pablo, Perú
https://dina.concytec.gob.pe/appDirectorioCTI/VerDatosInvestigador.do?id_investigador=39932
This research presents the problem of generating a humanoid gait from the imitation of a human being, using kinematic information to capture human movement and will apply a proposal based on learning by reinforcement in a continuous space. The proposal shows a scheme that takes the sequence of a human gait and the replica in a humanoid robot using a mapping algorithm. These results are not sufficient, since the replication of movements does not solve the problem of equilibrium. For this reason, an RBF-Sarsa (λ) algorithm is proposed which is a learning algorithm by reinforcement that uses possible actions at each step and interpolates them in this neural network. This network uses a reward function that is given by the torso deviation angle that measures the stability of the robot. The network stores knowledge that qualifies possible future actions by following a sequence of movements to achieve equilibrium. The results show valid solutions using the mapping algorithm, as well as RBF-Sarsa (λ) Algorithm convergence graphs.