Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 5
Encontrada(s) 1 página(s)

e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publication Contribución de congreso (ConferenceOutput)

Towards Unsupervised Knowledge Extraction

  • Tsatsou, Dorothea
  • Karangeorgos, Konstantinos
  • Dimou, Anastasios
  • Carbó Rubiera, Javier Ignacio
  • Molina López, José Manuel
  • Daras, Petros
22-24 marzo 2021, Integration of symbolic and sub-symbolic approaches is rapidly emerging as an Artificial Intelligence (AI) paradigm. This paper presents a proof-of-concept approach towards an unsupervised learning method, based on Restricted Boltzmann Machines (RBMs), for extracting semantic associations among prominent entities within data. Validation of the approach is performed in two datasets that connect language and vision, namely Visual Genome and GQA. A methodology to formally structure the extracted knowledge for subsequent use through reasoning engines is also offered., This work has been supported by the European Commission under Grant Agreement No. 787061
Repositorio Digital de la UPF
Publication Artículo científico (JournalArticle)

Complementary interactions between classical and top-down driven inhibitory mechanisms of attention

  • Low, Sock Ching
  • Vouloutsi, Vasiliki
  • Verschure, Paul F. M. J.
Selective attention informs decision-making by biasing perceptual processing towards task-relevant stimuli. In experimental and computational literature, this is most often implemented through top-down excitation of selected stimuli. However, physiological and anatomical evidence shows that in certain situations, top-down signals could instead be inhibitory. In this study, we investigated how such an inhibitory mechanism of top-down attention compares with an excitatory one. We did so in a neurorobotics context where the agent was controlled using an established hierarchical architecture. We augmented the architecture with an attentional system that implemented top-down attention biasing as connection gains. We tested four models of top-down attention on the simulated agent performing a foraging task: without top-down biasing, with only excitatory top-down gain, with only inhibitory top-down gain, and with both excitatory and inhibitory top-down gain. We manipulated the reward-distractor ratio that was presented and assessed the agent's performance using accumulated rewards and the latency of the selection. Using these measures, we provide evidence that excitatory and inhibitory mechanisms of attention complement each other., This project was funded by H2020 Research and Innovation program (#787061, ANITA), ERC H2020 (#840052, Cognitive RGS), H2020-Research and Innovation action EU.3.1.5.3 (#826421, VBC), and H2020-EU.2.1.5.1. (#820742, HR-Recycler). There are no conflicts of interest to declare.
Repositorio Digital de la UPF
Publication Contribución de congreso (ConferenceOutput)

A virtual reality system for the simulation of neurodiversity

  • López-Carral, Héctor
  • Blancas Muñoz, Maria
  • Mura, Anna
  • Omedas, Pedro
  • España-Cumellas, Àdria
  • Martínez Bueno, Enrique
  • Milliken, Neil
  • Moore, Paul
  • Haque, Leena
  • Gilroy, Sean
  • Verschure, Paul F. M. J.
Comunicació presentada a 6th International Congress on Information and Communication Technology (ICICT 2021), celebrat del 25 al 26 de febrer de 2021 de manera virtual., Autism is a neurodevelopmental disorder characterized by deficits in social communication and repetitive patterns of behavior. Individuals affected by Autism Spectrum Disorder (ASD) may face overwhelming sensory hypersensitivities that hamper their everyday life. In order to promote awareness about neurodiversity among the neurotypical population, we have developed an interactive virtual reality simulation to experience the oversensory stimulation that an individual with autism spectrum disorder may experience in a natural environment. In this experience, we project the user in a first-person perspective in a classroom where a teacher is presenting a lecture. As the user explores the classroom and attends the lecture, he/she is confronted with sensory distortions which are commonly experienced by persons with ASD. We provide the users with a virtual reality headset with motion tracking, two wireless controllers for interaction, and a wristband for physiological data acquisition to create a closed feedback loop. This wearable device measures blood volume pulse (BVP) and electrodermal activity (EDA), which we use to perform online estimations of the arousal levels of users as they respond to the virtual stimuli. We use this information to modulate the intensity of auditory and visual stimuli simulating a vicious cycle in which increased arousal translates into increased oversensory stimulation. Here, we present the architecture and technical implementation of this system., This research received funding from H2020-EU, ID: 787061 and ERC (PoC) H2020-EU, ID: 840052.
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Complementary interactions between classical and top-down driven inhibitory mechanisms of attention

  • Low, Sock Ching
  • Vouloutsi, Vasiliki
  • Verschure, Paul F. M. J.
Selective attention informs decision-making by biasing perceptual processing towards task-relevant stimuli. In experimental and computational literature, this is most often implemented through top-down excitation of selected stimuli. However, physiological and anatomical evidence shows that in certain situations, top-down signals could instead be inhibitory. In this study, we investigated how such an inhibitory mechanism of top-down attention compares with an excitatory one. We did so in a neurorobotics context where the agent was controlled using an established hierarchical architecture. We augmented the architecture with an attentional system that implemented top-down attention biasing as connection gains. We tested four models of top-down attention on the simulated agent performing a foraging task: without top-down biasing, with only excitatory top-down gain, with only inhibitory top-down gain, and with both excitatory and inhibitory top-down gain. We manipulated the reward-distractor ratio that was presented and assessed the agent's performance using accumulated rewards and the latency of the selection. Using these measures, we provide evidence that excitatory and inhibitory mechanisms of attention complement each other., This project was funded by H2020 Research and Innovation program (#787061, ANITA), ERC H2020 (#840052, Cognitive RGS), H2020-Research and Innovation action EU.3.1.5.3 (#826421, VBC), and H2020-EU.2.1.5.1. (#820742, HR-Recycler). There are no conflicts of interest to declare.
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

A virtual reality system for the simulation of neurodiversity

  • López-Carral, Héctor
  • Blancas Muñoz, Maria
  • Mura, Anna
  • Omedas, Pedro
  • España-Cumellas, Àdria
  • Martínez Bueno, Enrique
  • Milliken, Neil
  • Moore, Paul
  • Haque, Leena
  • Gilroy, Sean
  • Verschure, Paul F. M. J.
Comunicació presentada a 6th International Congress on Information and Communication Technology (ICICT 2021), celebrat del 25 al 26 de febrer de 2021 de manera virtual., Autism is a neurodevelopmental disorder characterized by deficits in social communication and repetitive patterns of behavior. Individuals affected by Autism Spectrum Disorder (ASD) may face overwhelming sensory hypersensitivities that hamper their everyday life. In order to promote awareness about neurodiversity among the neurotypical population, we have developed an interactive virtual reality simulation to experience the oversensory stimulation that an individual with autism spectrum disorder may experience in a natural environment. In this experience, we project the user in a first-person perspective in a classroom where a teacher is presenting a lecture. As the user explores the classroom and attends the lecture, he/she is confronted with sensory distortions which are commonly experienced by persons with ASD. We provide the users with a virtual reality headset with motion tracking, two wireless controllers for interaction, and a wristband for physiological data acquisition to create a closed feedback loop. This wearable device measures blood volume pulse (BVP) and electrodermal activity (EDA), which we use to perform online estimations of the arousal levels of users as they respond to the virtual stimuli. We use this information to modulate the intensity of auditory and visual stimuli simulating a vicious cycle in which increased arousal translates into increased oversensory stimulation. Here, we present the architecture and technical implementation of this system., This research received funding from H2020-EU, ID: 787061 and ERC (PoC) H2020-EU, ID: 840052.