TropNNC: Structured Neural Network Compression Using Tropical Geometry

cs.CV

Submitted: 2024-09-05

Updated: 2026-10-11

Comments: v4: expanded with Theorems 7 and 8, connection to SVD, argument for non-ReLU activations, and the discussion. Published in TMLR. Official record available at: https://openreview.net/forum?id=u7DRq1icmY

Journal ref: Published in Transactions on Machine Learning Research (06/2026)

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Sources

• A Tropical Approach to Neural Networks with Piecewise Linear Activations

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models