Wendler, Chris; Alistarh, Dan-AdrianIST Austria; Püschel, Markus
We present a novel class of convolutional neural networks (CNNs) for set functions,i.e., data indexed with the powerset of a finite set. The convolutions are derivedas linear, shift-equivariant functions for various notions of shifts on set functions.The framework is fundamentally different from graph convolutions based on theLaplacian, as it provides not one but several basic shifts, one for each element inthe ground set. Prototypical experiments with several set function classificationtasks on synthetic datasets and on datasets derived from real-world hypergraphsdemonstrate the potential of our new powerset CNNs.
NIPS: Conference on Neural Information Processing Systems
2019-12-08 – 2019-12-14
Wendler C, Alistarh D-A, Püschel M. Powerset convolutional neural networks. In: Vol 32. Neural Information Processing Systems Foundation; 2019:927-938.
Wendler, C., Alistarh, D.-A., & Püschel, M. (2019). Powerset convolutional neural networks (Vol. 32, pp. 927–938). Presented at the NIPS: Conference on Neural Information Processing Systems, Vancouver, Canada: Neural Information Processing Systems Foundation.
Wendler, Chris, Dan-Adrian Alistarh, and Markus Püschel. “Powerset Convolutional Neural Networks,” 32:927–38. Neural Information Processing Systems Foundation, 2019.
C. Wendler, D.-A. Alistarh, and M. Püschel, “Powerset convolutional neural networks,” presented at the NIPS: Conference on Neural Information Processing Systems, Vancouver, Canada, 2019, vol. 32, pp. 927–938.
Wendler C, Alistarh D-A, Püschel M. 2019. Powerset convolutional neural networks. NIPS: Conference on Neural Information Processing Systems vol. 32. 927–938.
Wendler, Chris, et al. Powerset Convolutional Neural Networks. Vol. 32, Neural Information Processing Systems Foundation, 2019, pp. 927–38.