Now showing items 1-20 of 36

    • An application of explainability methods in reinforcement learning 

      Climent Muñoz, Antoni (Universitat Politècnica de Catalunya, 2020-07-02)
      Bachelor thesis
      Open Access
      La popularidad de los métodos explicativos está aumentando en el contexto de la Inteligencia Artificial y consiste en dar explicaciones inteligibles a modelos complejos. Recientemente, en el contexto del Aprendizaje Reforzado ...
    • Applying and verifying an explainability method based on policy graphs in the context of reinforcement learning 

      Climent Muñoz, Antoni; Gnatyshak, Dmitry; Álvarez Napagao, Sergio (IOS Press, 2021)
      Conference report
      Open Access
      The advancement on explainability techniques is quite relevant in the field of Reinforcement Learning (RL) and its applications can be beneficial for the development of intelligent agents that are understandable by humans ...
    • Applying multi-agent reinforcement learning to solve sequential moral dilemmas 

      Choinski, Michal (Universitat Politècnica de Catalunya, 2021-04-26)
      Master thesis
      Open Access
      Covenantee:   Universitat de Barcelona. Facultat de Matemàtiques i Informàtica / Universitat Rovira i Virgili
      Incorporation of ethical values in the field of Artificial Intelligence is inevitable. With the rapid development of technologies capable of making autonomous decisions, more attention should be dedicated to the process ...
    • Applying the rainbow architecture to intrusion detection systems 

      Izquierdo García-Faria, Tomás (Universitat Politècnica de Catalunya, 2021-04-28)
      Master thesis
      Open Access
      There is a lot of expectation on how Artificial Intelligence (AI) is going to have an impact on Cybersecurity. From new sophisticated attacks to new ways of defending a system from cybercriminals. A lot of techniques are ...
    • Aprendizaje por refuerzo aplicado a los videojuegos cooperativos 

      Alcocer Soto, Daniel (Universitat Politècnica de Catalunya, 2018-07)
      Bachelor thesis
      Open Access
      Se ha desarrollado un algoritmo que combina el uso de redes neuronales con el algoritmo Q-learning para aprender a jugar con la ayuda de un jugador humano y aprender a cooperar con él para ganar en un videojuego basado en ...
    • Aprendizaje por refuerzo aplicado a personajes no controlables en Minetest 

      Romero Reviriego, Aitor (Universitat Politècnica de Catalunya, 2019-01)
      Bachelor thesis
      Open Access
      En este proyecto se ha utilizado la rama de la IA llamada aprendizaje por refuerzo y intenta desarrollar agentes para el videojuego Minetest que actúen como aliados del jugador.
    • Aprenentatge aplicat al NL Texas Hold'em 

      Perapoch Amadó, Marçal (Universitat Politècnica de Catalunya, 2014-06-06)
      Master thesis (pre-Bologna period)
      Open Access
      Aquest projecte consisteix en la construcció d'un sistema software que permeti observar el comportament i resultats d'aplicar tècniques d'Intel·ligència Artificial al domini que engloba el conegut joc de cartes: Pòquer.
    • Combining long-short term memory and reinforcement learning for improved autonomous network operation 

      Tabatabaeimehr, Fatemehsadat; Barzegar, Sima; Ruiz Ramírez, Marc; Velasco Esteban, Luis Domingo (Institute of Electrical and Electronics Engineers (IEEE), 2021)
      Conference report
      Open Access
      A combined LSTM and RL approach is proposed for dynamic connection capacity allocation. The LSTM predictor anticipates periodical long-term sharp traffic changes and extends short-term RL knowledge. Numerical results show ...
    • Deep reinforcement learning IA para Starcraft 2 

      Roldán Montaner, Carlos (Universitat Politècnica de Catalunya, 2018-06)
      Bachelor thesis
      Open Access
      Este proyecto intenta desarrollar agentes de aprendizaje por refuerzo para escenarios concretos de Starcraft 2 y sacar conclusiones sobre cuales son las dificultades más importantes y los enfoques más adecuados para afrontar ...
    • Desarrollo de un bot para un juego de lucha mediante aprendizaje por refuerzo 

      Balaghi Buil, David (Universitat Politècnica de Catalunya, 2018-10-23)
      Bachelor thesis
      Open Access
      Este proyecto trata el diseño e implementación de un videojuego de lucha en 2D desarrollado en Unity, y de un bot que aprende y mejora a medida que lo juega mediante una IA que implementa técnicas de aprendizaje por refuerzo.
    • Distributed Deep Reinforcement Learning in an HPC system and deployment to the Cloud 

      Escobar Castells, Miquel (Universitat Politècnica de Catalunya, 2021-07-01)
      Bachelor thesis
      Open Access
      Combinar l'aprenentatge per reforç amb l'aprenentatge profund és, a dia d'avui, un dels reptes més grans en el sector d'investigació en intel·ligència artificial. Escalar aquest tipus d'aplicacions mitjançant supercomputadors ...
    • Evaluating the Robustness of GAN-Based Inverse Reinforcement Learning Algorithms 

      Heidecke, Johannes (Universitat Politècnica de Catalunya, 2019-01-15)
      Master thesis
      Open Access
      Covenantee:   Universitat de Barcelona / Universitat Rovira i Virgili
      We evaluate the robustness of reward functions learned with IRL, when transferred to similar tasks. We exceed state of the art results for one benchmark task and solve another one for the first time. Modifications are ...
    • Evolving cooperation in multi-agent systems 

      Elli Galata, Aglaia (Universitat Politècnica de Catalunya, 2021-04-28)
      Master thesis
      Open Access
      Covenantee:   Universitat de Barcelona / Universitat Rovira i Virgili
      Applications of deep reinforcement learning in multi-agent systems are a rapidly developing scientific field that focuses on constructing computational frameworks and models for how various agents can interact efficiently ...
    • Extending Markov games to learn policies aligned with moral values 

      Rodríguez Soto, Manel (Universitat Politècnica de Catalunya, 2019-04)
      Master thesis
      Restricted access - confidentiality agreement
    • From one to many: Simulating groups of agents with reinforcement learning controllers 

      Casadiego, Luiselena; Pelechano Gómez, Núria (2015)
      Conference lecture
      Open Access
      Simulation of crowd behavior has been approached through many different methodologies, but the problem of mimicking human decisions and reactions remains a challenge for all. We propose an alternative model for simulation ...
    • How could fish win a race through Reinforcement Learning 

      Sánchez Molina, David (Universitat Politècnica de Catalunya, 2020-06-30)
      Bachelor thesis
      Open Access
      Covenantee:   University of Colorado Colorado Springs
      Fish generate propulsion through body and caudal (tail) fin undulation. The undulation kinematics, suchas the amplitude and the frequency, determines the generated hydrodynamic force, associated accelera-tion, and therefore ...
    • Learning complex games through self play - Pokémon battles 

      Llobet Sanchez, Miquel (Universitat Politècnica de Catalunya, 2018-06)
      Bachelor thesis
      Open Access
      En aquest projecte s'analitza la viabilitat d'utilitzar aprenentatge per reforç i "self- play" per entrenar un agent a jugar Batalles Pokémon. El joc és analitzat en detall i les seves propietats úniques són revelades. El ...
    • Learning recursive goal proposal: a hierarchical reinforcement learning approach 

      Palliser Sans, Rafel (Universitat Politècnica de Catalunya, 2021-04-28)
      Master thesis
      Open Access
      Covenantee:   Universitat de Barcelona / Universitat Rovira i Virgili. Escola Tècnica Superior d'Enginyeria
      Reinforcement Learning's unique way of learning has led to remarkable successes like Alpha Zero or Alpha Go, mastering the games of Chess and Go, and being able to beat the respective World Champions. Notwithstanding, ...
    • Learning to run naturally: guiding policies with the Spring-Loaded Inverted Pendulum 

      Ordoñez Apraez, Daniel Felipe (Universitat Politècnica de Catalunya, 2021-04-28)
      Master thesis
      Open Access
      Covenantee:   Universitat de Barcelona / Universitat Rovira i Virgili
      In this work, we proposed a new approach for learning legged locomotion for any legged robot, in the sagittal plane, by using a combination of classical control techniques and reinforcement learning. Specifically, we use ...
    • Learning to safely drive using Reinforcement Learning 

      Carrera Escalé, Laura (Universitat Politècnica de Catalunya, 2021-04-28)
      Master thesis
      Open Access
      Covenantee:   Universitat de Barcelona / Universitat Rovira i Virgili
      The autonomous driving research area has gained popularity over the past decade, even more with the launch of the first autonomous vehicle from Tesla, Inc. Different research branches are currently being studied, and one ...