Research map / Federated learning and privacy / Communication-efficient FL
DeepAFL: Deep Analytic Federated Learning
Jianheng Tang, Yajiang Huang, Kejia Fan, Feijiang Han, Jiaxu Li, Jinfeng Xu, Run He, Anfeng Liu, Houbing Herbert Song, Huiping Zhuang, Yunhuai Liu
This paper tried a common change: Add residual connections
add residual skip connection to deep analytic layers · CIFAR-100, Tiny-ImageNet, CIFAR-10 · accuracy vs layers T Improved
What 546 accepted papers report for this change
- 449 improved
- 35 no clear effect
- 11 mixed
- 51 hurt
- Resolution as a Direction: Vector-Panning Feature Alignment for Cross-Resolution Re-Identification ICML 2026 · add residual connection to VP · MLR-Market-1501 · Rank-1 (%): 91.2 -> 94.1Improved
- MAGIC: Multi-Granularity Language-Informed Image Clustering ICML 2026 · add residual connection to semantic adapters · CIFAR-20 · ACC: 66.1 -> 59.9Hurt
- Domain Adaptive Object Detection via Dynamic Causal Refinement ICML 2026 · add residual connection to SCC · Cityscapes to Foggy Cityscapes (0.02 split) · mAP@50: 64.0 -> 64.6Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 |
|---|---|---|---|
| add residual skip connection to deep analytic layers (ablation) | DeepAFL · CIFAR-100, Tiny-ImageNet, CIFAR-10 | accuracy vs layers T | Improved |
| add per-layer random projection B_t instead of identity (ablation) | DeepAFL · CIFAR-100, Tiny-ImageNet, CIFAR-10 | accuracy | Improved |
| add activation function in residual block (ablation) | DeepAFL · CIFAR-100, Tiny-ImageNet, CIFAR-10 | accuracy | Improved |
| add trainable transformation Omega_{t+1} in residual block (ablation) | DeepAFL · CIFAR-100, Tiny-ImageNet, CIFAR-10 | accuracy | Improved |
| stack analytic layers with random projections only (no activation) | naive deepening · CIFAR-100 | accuracy | No clear effect |
| add activation after random projection in stacked layers (no residual) | naive deepening · CIFAR-100 | accuracy vs layers | Mixed |
| add activated random projection to form zero-layer features Phi_0 | AFL · CIFAR-100 | accuracy: about 2.1% | Improved |
| increase layers T from 5 to 20 | DeepAFL · CIFAR-100 | top-1 accuracy (%): 64.72 -> 66.98 | Improved |
| increase layers T from 5 to 20 | DeepAFL · Tiny-ImageNet | top-1 accuracy (%): 60.31 -> 62.35 | Improved |
| increase layers T from 20 to 50 | DeepAFL · CIFAR-100 and Tiny-ImageNet | testing accuracy: over 1.5% | Improved |
| use GELU instead of weaker activations (Softshrink, none) (ablation) | DeepAFL · activation study | accuracy: about 2% | Improved |
| increase projection dimensions d_Phi and d_F to 2^13 | DeepAFL · dimension study | numerical stability | Hurt |
Closest papers by meaning
- AFL: A Single-Round Analytic Approach for Federated Learning with Pre-trained Models (CVPR 2025 · similarity 0.91)
- FedQS: Optimizing Gradient and Model Aggregation for Semi-Asynchronous Federated Learning (NeurIPS 2025 · similarity 0.87)
- An Aggregation-Free Federated Learning for Tackling Data Heterogeneity (CVPR 2024 · similarity 0.87)
- Convergence Analysis of Split Federated Learning on Heterogeneous Data (NeurIPS 2024 · similarity 0.86)
- Revisiting Ensembling in One-Shot Federated Learning (NeurIPS 2024 · similarity 0.86)
- FedDAT: An Approach for Foundation Model Finetuning in Multi-Modal Heterogeneous Federated Learning (AAAI 2024 · similarity 0.86)
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.
