Machine learning via dynamical processes in complex networks
Resumo
Machine learning is a research field devoted to the development of techniques capable of enabling a machine to ”learn” from data. Many techniques have been proposed so far, but there are still issues to be unveiled specially in interdisciplinary research. In the doctorate thesis summarized here, we explore the advantages of network data representation to develop learning techniques based on dynamical processes. Our studies covered the three main learning categories: supervised, semi-supervised and unsupervised. We also applied the developed techniques for images, handwritten digits and letters recognition, and many other classification tasks. Advantages of network data representation and dynamical processes for machine learning are highlighted in all cases.
Referências
[Breve et al. 2012] Breve, F., Zhao, L., Quiles, M. G., Pedrycz,W., and Liu, J. (2012). Particle competition and cooperation in networks for semi-supervised learning. Knowledge and Data Engineering, IEEE Transactions on, 24(9):1686–1698.
[Silva et al. 2013] Silva, T. C., Zhao, L., and Cupertino, T. H. (2013). Handwritten data clustering using agents competition in networks. Journal of Mathematical Imaging and Vision, 45(3):264–276.
[Yan et al. 2007] Yan, S., Xu, D., Zhang, B., Zhang, H.-J., Yang, Q., and Lin, S. (2007). Graph embedding and extensions: A general framework for dimensionality reduction. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 29(1):40–51.
