Description
| Title | Convolutional neural networks for knowledge graphs. |
| Abstract | Graph convolutional networks are a new model for the analysis of graphs, inspired by the convolutional networks currently being used to solve everything from image analysis to Go. Together with the AMLab at the UvA, we have been developing an extension of existing graph convolutional neural networks to knowledge graphs. I'll be explaining the principle, and showing some early results. |
Other presentations by Peter Bloem
| Date | Title |
|---|---|
| 25 April 2016 | Learning fractals by Expectation-Maximization |
| 29 August 2016 | Network motif detection at scale |
| 27 February 2017 | Convolutional neural networks for knowledge graphs. |
| 19 March 2018 | Adaptive, sparse hypernetworks: Learning sparse, adaptive layers through backpropagation. |
| 06 May 2019 | Differentiable quicksort |