Research map / Label-efficient and robust learning / Recognition and long-tail
I Am Going MAD: Maximum Discrepancy Competition for Comparing Classifiers Adaptively
Haotao Wang, Tianlong Chen, Zhangyang Wang, Kede Ma
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 |
|---|---|---|---|
| Use MAD-selected images to compare models rather than relying only on the fixed ImageNet validation ranking | ImageNet classifiers · ImageNet validation versus MAD test set | Spearman rank-order correlation coefficient: 0.89 | Mixed |
| Add squeeze-and-excitation channel-dependency modeling | ResNet-101 · MAD test set | rank: up by two places | Improved |
| Use progressive cell-wise architecture search instead of global architecture search | NASNet-A-Large · MAD test set | rank: 4 -> 7 | Hurt |
| Use weakly supervised Instagram hashtag pretraining | WSL-ResNeXt101-32×48 · MAD test set | rank: 1 -> 2 | Improved |
| Use compound scaling | EfficientNet-B7 · MAD test set | rank: 2 -> 1 | Improved |
| Increase the number of selected images per classifier pair beyond 15 (ablation) | MAD · MAD image-classifier ranking | SRCC versus top-30 ranking: >0.90 | Improved |
Closest papers by meaning
- L-WISE: Boosting Human Visual Category Learning Through Model-Based Image Selection and Enhancement (ICLR 2025 · similarity 0.88)
- Beyond accuracy: quantifying trial-by-trial behaviour of CNNs and humans by measuring error consistency (NeurIPS 2020 · similarity 0.87)
- Enriching ImageNet With Human Similarity Judgments and Psychological Embeddings (CVPR 2021 · similarity 0.87)
- Causal Transportability for Visual Recognition (CVPR 2022 · similarity 0.87)
- Intriguing Properties of Generative Classifiers (ICLR 2024 · similarity 0.86)
- OddGridBench: Exposing the Lack of Fine-Grained Visual Discrepancy Sensitivity in Multimodal Large Language Models (CVPR 2026 · 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.
