Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Bayesian Sparsification of Deep C-valued Networks
Ivan Nazarov, Evgeny Burnaev
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Fine-tuning after C-valued variational sparsification (VD/ARD) at high compression | C-valued DeepConvNet on MusicNet · MusicNet | average precision | Improved |
| Sparsify with C-valued VD or ARD and compress CVNN | C-valued DeepConvNet of Trabelsi et al. · MusicNet | compression at comparable AP | Improved |
| ARD prior instead of VD for C-valued sparsification | C-valued networks · MusicNet and MNIST-like | average precision and compression | Mixed |
| Halving C-valued network width vs doubling R-valued network width | MNIST-like networks · MNIST (fft/raw) | compression and accuracy | Hurt |
| Fourier-transform input features vs raw input | MNIST-like networks · MNIST | accuracy at high compression | Mixed |
| Lower sparsity coefficient C in KL term (undercompressed models) with fine-tuning | C-valued DeepConvNet on MusicNet · MusicNet | early stopping epoch | Hurt |
| Pruning threshold tau variation during fine-tuning only (ablation) | C-valued DeepConvNet on MusicNet · MusicNet | performance-compression curve | No clear effect |
| Halve receptive field of first convolution (k3) in Trabelsi network, with C-VD (ablation) | C-valued DeepConvNet on MusicNet · MusicNet | average precision at x100 compression | Mixed |
Closest papers by meaning
- Deep Complex Networks (ICLR 2018 · similarity 0.86)
- Variational Dropout Sparsifies Deep Neural Networks (ICML 2017 · similarity 0.86)
- Efficient Variational Inference for Sparse Deep Learning with Theoretical Guarantee (NeurIPS 2020 · similarity 0.85)
- Training Bayesian Neural Networks with Sparse Subspace Variational Inference (ICLR 2024 · similarity 0.84)
- Sparse Bayesian Generative Modeling for Compressive Sensing (NeurIPS 2024 · similarity 0.83)
- Random Sparse Lifts: Construction, Analysis and Convergence of finite sparse networks (ICLR 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
