Research map / Architectures and efficiency / RNNs and normalization
Learning Time/Memory-Efficient Deep Architectures With Budgeted Super 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 |
|---|---|---|---|
| Search for an architecture under an inference-computation budget instead of using a fixed ResNet | ResNet Fabric on CIFAR-10 · CIFAR-10 | accuracy (%): 92.19 -> 92.39 | Helped |
| Search for an architecture under a computation budget instead of using a fixed CNF | CNF on CIFAR-10 · CIFAR-10 | accuracy (%): 92.54 -> 93.14 | Helped |
| Search for an architecture under a memory budget instead of using a fixed ResNet | ResNet Fabric on CIFAR-10 · CIFAR-10 | classification error: 7.81% -> 6.58% | Helped |
| Search for an architecture under a computation budget instead of using full CNF | CNF on Part Label · Part Label | error rate: 4.94% -> 4.57% | Helped |
| Use a budgeted architecture for distributed computation with 4 cores | ResNet Fabric on CIFAR-10 · CIFAR-10, n=4 | accuracy (%): 93.20 -> 94.36; 94.09 -> 94.36 | Mixed |
| Use a budgeted architecture for distributed computation with 1 core | CNF on CIFAR-10 · CIFAR-10, n=1 | accuracy (%): 94.83 -> 95.02 | Helped |
| Increase distributed core count to n=4, making the cost constraint less regularizing | Budgeted architectures · CIFAR-10 | accuracy | Hurt |
Closest papers by meaning
- Budgeted Training: Rethinking Deep Neural Network Training Under Resource Constraints (ICLR 2020 · similarity 0.89)
- GreedyNAS: Towards Fast One-Shot NAS With Greedy Supernet (CVPR 2020 · similarity 0.87)
- Efficient Few-Shot Neural Architecture Search by Counting the Number of Nonlinear Functions (AAAI 2025 · similarity 0.86)
- Subnet-Aware Dynamic Supernet Training for Neural Architecture Search (CVPR 2025 · similarity 0.85)
- Revisiting Neural Networks for Few-Shot Learning: A Zero-Cost NAS Perspective (ICML 2025 · similarity 0.85)
- MEET: Towards Memory-Efficient Temporal Sparse Deep Neural Networks (CVPR 2025 · similarity 0.85)
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.