Multi-layered tensor networks for image classification
Research output: Contribution to conference › Paper › Research
Standard
Multi-layered tensor networks for image classification. / Selvan, Raghavendra; Ørting, Silas; Dam, Erik B.
2020. Paper presented at 1st Workshop on Quantum Tensor Networks in Machine Learning, Online.Research output: Contribution to conference › Paper › Research
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - CONF
T1 - Multi-layered tensor networks for image classification
AU - Selvan, Raghavendra
AU - Ørting, Silas
AU - Dam, Erik B
N1 - Accepted to the First Workshop on Quantum Tensor Networks in Machine Learning. In conjunction with 34th NeurIPS, 2020. Source code at https://github.com/raghavian/mltn
PY - 2020
Y1 - 2020
N2 - The recently introduced locally orderless tensor network (LoTeNet) for supervised image classification uses matrix product state (MPS) operations on grids of transformed image patches. The resulting patch representations are combined back together into the image space and aggregated hierarchically using multiple MPS blocks per layer to obtain the final decision rules. In this work, we propose a non-patch based modification to LoTeNet that performs one MPS operation per layer, instead of several patch-level operations. The spatial information in the input images to MPS blocks at each layer is squeezed into the feature dimension, similar to LoTeNet, to maximise retained spatial correlation between pixels when images are flattened into 1D vectors. The proposed multi-layered tensor network (MLTN) is capable of learning linear decision boundaries in high dimensional spaces in a multi-layered setting, which results in a reduction in the computation cost compared to LoTeNet without any degradation in performance.
AB - The recently introduced locally orderless tensor network (LoTeNet) for supervised image classification uses matrix product state (MPS) operations on grids of transformed image patches. The resulting patch representations are combined back together into the image space and aggregated hierarchically using multiple MPS blocks per layer to obtain the final decision rules. In this work, we propose a non-patch based modification to LoTeNet that performs one MPS operation per layer, instead of several patch-level operations. The spatial information in the input images to MPS blocks at each layer is squeezed into the feature dimension, similar to LoTeNet, to maximise retained spatial correlation between pixels when images are flattened into 1D vectors. The proposed multi-layered tensor network (MLTN) is capable of learning linear decision boundaries in high dimensional spaces in a multi-layered setting, which results in a reduction in the computation cost compared to LoTeNet without any degradation in performance.
KW - cs.CV
KW - cs.LG
KW - stat.ML
M3 - Paper
T2 - 1st Workshop on Quantum Tensor Networks in Machine Learning
Y2 - 11 December 2020
ER -
ID: 255889452