Research map / Learning theory and algorithms / Matrix and tensor methods
Tensor Switching Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace the SS-ReLU kernel with the TS-ReLU kernel | SVM · SVHN2 | error rate (%): 21.60 -> 20.38 | Helped |
| Replace the SS-ReLU kernel with the TS-ReLU kernel | SVM · MNIST | error rate (%): 1.40 -> 1.40 | No clear effect |
| Replace the SS-ReLU kernel with the TS-ReLU kernel | SVM · CIFAR10 | error rate (%): 43.18 -> 43.60 | Hurt |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | MLP · MNIST, asymptotic | error rate (%): 2.36 -> 2.06 | Helped |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | MLP · CIFAR10, asymptotic | error rate (%): 46.91 -> 46.87 | No clear effect |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | MLP · SVHN2, asymptotic | error rate (%): 12.20 -> 12.58 | Hurt |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | CNN · MNIST, asymptotic | error rate (%): 1.08 -> 0.86 | Helped |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | CNN · CIFAR10, asymptotic | error rate (%): 26.73 -> 25.74 | Helped |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | CNN · SVHN2, asymptotic | error rate (%): 4.96 -> 5.06 | Hurt |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | CNN · MNIST, one-pass | error rate (%): 43.74 -> 3.85 | Helped |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | CNN · CIFAR10, one-pass | error rate (%): 74.84 -> 54.40 | Helped |
| Use TS nonlinearities with linear rotation-compression instead of SS nonlinearities | CNN · SVHN2, one-pass | error rate (%): 13.69 -> 9.13 | Helped |
Closest papers by meaning
- Convolutions and More as Einsum: A Tensor Network Perspective with Advances for Second-Order Methods (NeurIPS 2024 · similarity 0.87)
- Factor Augmented Tensor-on-Tensor Neural Networks (AAAI 2025 · similarity 0.84)
- Global Convergence and Rich Feature Learning in $L$-Layer Infinite-Width Neural Networks under $\mu$ Parametrization (ICML 2025 · similarity 0.84)
- Better NTK Conditioning: A Free Lunch from (ReLU) Nonlinear Activation in Wide Neural Networks (NeurIPS 2025 · similarity 0.83)
- Deep Tree Tensor Networks (NeurIPS 2025 · similarity 0.83)
- tnGPS: Discovering Unknown Tensor Network Structure Search Algorithms via Large Language Models (LLMs) (ICML 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.