Facultad Regional Buenos Aires

Permanent URI for this communityhttp://48.217.138.120/handle/20.500.12272/6

Browse

Subject: search.filters.subject.machine learning

Search Results

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Moving in a Simulated Environment Through Deep Reinforcement Learning
    (IEEE, 2022-09-08) Esarte, Javier; Folino, Pablo Daniel; Gómez, Juan Carlos
    Reinforcement learning is a field of artificial intelligence that is continuously evolving and has a wide variety of applications. In recent years major progress has been made in the application of deep reinforcement learning to highdimensional problems with continuous state and action spaces. This paper presents a complete analysis of the application of the soft actor-critic algorithm to teach a four legged robot with three joints on each leg how to move towards the center of a virtually simulated environment. The general formulation of the reinforcement learning problem is first presented, followed by the description of the environment under analysis and the applied algorithm. Afterwards, the obtained results are compared against those of a manually programmed policy, closing with a discussion of some key design choices and common challenges.
  • Thumbnail Image
    Item
    Evaluation of the bias in the management of patient’s appointments in a pediatric office
    (2020-04-01) Vegega, Cinthia; Pytel, Pablo; Pollo Cattaneo, María Florencia
    The application of Machine Learning algorithms must always take into account the objectives set within the project, the characteristics of the domain where the project will be carried out and the data available to use. Given this, it is essential before collecting data considered as representative of the problem to be solved, because otherwise there may be hidden biases in the data and these may solve a different problem from the one intended. In this context, the aim of this work is to apply a process based on the Gridding method that allows the analysis of the features of the data to be used. This process is applied to the historical data of a pediatric medical office where it is sought to implement an intelligent system that allows to predict the number of normal and overshift appointments for a particular date and time, since it is desired to hire, when necessary, another pediatric doctor to assist in the care of patients.
  • Thumbnail Image
    Item
    Predictor de deserción universitaria
    (2021-04-01) Romero, Giselle; Toranzo Calderón, Joaquin; Jaremczuk, Sebastián; Gómez, Juan Carlos; Verrastro, Claudio
    La deserción estudiantil siempre ha sido un tema de preocupación debido a sus múltiples implicancias. En este trabajo se propone la aplicación de técnicas de reconocimiento de patrones para exponer información útil y formular reglas de inferencia en sistemas de diagnóstico automático. De esta manera se generan modelos predictivos de deserción universitaria en la UTN.BA, a partir de bases de datos de estudiantes de la carrera de Ingeniería en Sistemas de la Información del plan K08. Se construyeron dos modelos, uno basado sobre Máquinas de Vectores de Soporte y otro sobre Redes neuronales. Ambos presentan resultados muy similares reconociendo a estudiantes en situación de deserción con una exactitud de 79%.