Research map / Architectures and efficiency / CNN design
Consensus vs. Controversy: Mapping the Decision Space Where Architectures Diverge
Minhyeok Lee
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 |
|---|---|---|---|
| apply isotonic calibration to model confidences (ablation) | 12 pretrained models · ImageNet val | Controversy Score: 4.46 -> 18.65 | Improved |
| apply isotonic calibration (ablation) | 12 pretrained models · ImageNet val | mean ECE: 0.089 -> 0.044 | Improved |
| raise gating escalation threshold from 0.50 to 0.90 (ablation) | ConvNeXt-Tiny -> ResNet-50 -> ViT-Base · ImageNet val | accuracy: 0.837 -> 0.825 | Hurt |
| raise gating escalation threshold from 0.50 to 0.90 (ablation) | sequential gating · ImageNet val | avg model calls: 1.18 -> 1.96 | Hurt |
| use more models to approximate controversy set (3 to 8) (ablation) | model subsets · ImageNet val | Jaccard overlap: 31.7% -> 57.1% | Improved |
| vary alpha threshold from 5% to 20% (ablation) | controversy split · ImageNet val | Controversy Score: 4.4 to 4.8 | No clear effect |
| use top-5 Jaccard instead of confidence std as disagreement (ablation) | controversy metric · ImageNet val | Controversy Score: 4.46 vs 3.05 | Mixed |
| evaluate under distribution shift | controversy framework · ImageNet-A / ImageNet-R | Controversy Score: 4.46 -> 2.8929 / 3.3384 | Mixed |
Closest papers by meaning
- Beyond accuracy: quantifying trial-by-trial behaviour of CNNs and humans by measuring error consistency (NeurIPS 2020 · similarity 0.87)
- Directional Confusions Reveal Divergent Inductive Biases Through Rate-Distortion Geometry in Human and Machine Vision (NeurIPS 2026 · similarity 0.85)
- Evaluating Multiview Object Consistency in Humans and Image Models (NeurIPS 2024 · similarity 0.85)
- Semi-Truths: A Large-Scale Dataset of AI-Augmented Images for Evaluating Robustness of AI-Generated Image detectors (NeurIPS 2024 · similarity 0.84)
- Comparing the Decision-Making Mechanisms by Transformers and CNNs via Explanation Methods (CVPR 2024 · similarity 0.84)
- LAION-C: An Out-of-Distribution Benchmark for Web-Scale Vision Models (ICML 2025 · similarity 0.84)
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.
