Aplicación de Deep Learning al aprendizaje de modelos en robótica cognitiva
- Francisco Bellas-Bouza
- Jose Becerra-Permuy
- Ariel Rodríguez-Jiménez
- Esteban Arias-Méndez
ISSN: 0379-3982, 2215-3241
Year of publication: 2020
Issue Title: Movilidad Estudiantil 7
Volume: 33
Issue: 6
Pages: 92-104
Type: Article
More publications in: Tecnología en Marcha
Abstract
The kind of training used for an artificial neural network will depend on factors such as: available data, training time, hardware resources, etc. The trainings can be online and offline. In the current article we experimented with online trainings on a robot whose main characteristic is the usage of a Cognitive Darwinist Mechanism to survive.The robot learns in real-time. It has deep artificial neural networks to predict actions, it’s trained using the least amount of storage and the training time has to be as fast as possible; keeping high confidence in the artificial neural network.The experimental trainings are: Online Deep Learning, Online Deep Learning with memory and Online Mini-Batch Deep Learning with memory
Bibliographic References
- Arrabales, R. (2007). Robótica Cognitiva (1era edición) [Online]. Disponible: http://www.conscious-robots.com/es/2007/08/21/robotica-cognitiva/
- Bottou, U. (2010). Large-Scale Machine Learning with Stochastic Gradient Descent (1era edición) [Online]. Disponible: https://leon.bottou.org/publications/pdf/compstat-2010.pdf
- Brownlee, J. (2017). A Gentle Introduction to Mini-Batch Gradient Descent and How to Configure Batch Size (1era edición) [Online]. Disponible: https://machinelearningmastery.com/gentle-introduction-mini-batch-gra-dient-descent-configure-batch-size/
- Crowford, C. (2016). An introduction to deep learning (1era edición) [Online]. Disponible: https://blog.algorith-mia.com/introduction-to-deep-learning/
- Duchi, S., Hazan, E. y Singer, Y. (2011). Adaptive Subgradient Method for Online Learning and Stochastic Optimization (1era edición) [Online]. Disponible: http://www.jmlr.org/papers/volume12/duchi11a/duchi11a.pdf
- F. Bellas, “MDB mecanismo cognitivo darwinista para agentes autónomos”, Disertación doctoral, departamen-to de computación, Universidade da Coruña, A Coruña, España, 2003.
- Guizzo, B. y Ackerman, E. (2018, setiembre 18). How Rethink Robotics Built its New Baxter Robot Worker (1era edición) [Online]. Available: https://spectrum.ieee.org/robotics/industrial-robots/rethink-robotics-baxter-robot-factory-worker
- Kingma, D., y Lei J. (2015). ADAM: A method for stochastic optimization (1era edición) [Online]. Disponible: https://arxiv.org/pdf/1412.6980.pdf
- Poggio, T., Voinea, S., y Rosasco, L. (2018). Online learning, stability and stochastic gradient descent (1era edición) [Online]. Disponible: https://arxiv.org/pdf/1105.4701.pdf
- Sahoo, D., Pham, Q., Lu J., y C.H S. (2017). Online Deep Learning: Learning Deep Neural Networks on the Fly (1era edición) [Online]. Disponible: https://arxiv.org/pdf/1711.03705.pdf
- Stanley, K. y Miikkulainen, R. (2002). Evolving Neural Networks Through Augmenting Topologies (1era edición) [Online]. Disponible: http://nn.cs.utexas.edu/downloads/papers/stanley.ec02.pdf[
- Zulkifli, H. (2018). Understanding Learning Rates and How It Improves Performance in Deep Learning (1era edición) [Online]. Disponible: https://towardsdatascience.com/understanding-learning-rates-and-how-it-impro-ves-performance-in-deep-learning-d0d4059c1c10