Research map / Label-efficient and robust learning / Recognition and long-tail
Inter-Class Angular Loss for Convolutional Neural 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 |
|---|---|---|---|
| add ICAL loss to softmax | ResNet-110 · CIFAR-100 | accuracy (%): 68.82 -> 70.48 | Helped |
| add ICAL loss to softmax | ResNet-110 · CIFAR-10 | accuracy (%): 90.31 -> 90.98 | Helped |
| add ICAL loss to softmax | WRN-28-10 · CIFAR-100 | accuracy (%): 75.77 -> 76.74 | Helped |
| add ICAL loss to softmax | DenseNet-BC-100 · CIFAR-100 with augmentation | accuracy (%): 77.73 -> 77.84 | No clear effect |
| add ICAL loss to softmax | DenseNet-BC-100 · CIFAR-10 with augmentation | accuracy (%): 95.49 -> 95.51 | No clear effect |
| add ICAL loss to softmax | ResNeXt-29 8x64d · CIFAR-10 | accuracy (%): 92.36 -> 92.68 | Helped |
| add ICAL loss to softmax | CNN-sen-rand · MPQA | accuracy (%): 83.4 -> 85.2 | Helped |
| add ICAL loss to softmax | LSTM-RNN · MR | accuracy (%): 77.4 -> 78.2 | Helped |
| add ICAL loss to softmax | CNN-sen-static · TREC | accuracy (%): 92.8 -> 93.3 | Helped |
| replace L-Softmax (m=4) with ICAL (λ=1.6) | Liu et al. network · CIFAR-10 | accuracy (%): 92.42 -> 93.06 | Helped |
| replace L-Softmax (m=4) with ICAL (λ=1.6) | Liu et al. network · CIFAR-100 | accuracy (%): 70.47 -> 71.09 | Helped |
| vary λ from 1.4 to 2.0 for ICAL (ablation) | Liu et al. network · CIFAR-10 | accuracy (%): 92.99 to 93.06 | No clear effect |
Closest papers by meaning
- OLÉ: Orthogonal Low-Rank Embedding - A Plug and Play Geometric Loss for Deep Learning (CVPR 2018 · similarity 0.86)
- On the Learning Property of Logistic and Softmax Losses for Deep Neural Networks (AAAI 2020 · similarity 0.85)
- PICNN: A Pathway towards Interpretable Convolutional Neural Networks (AAAI 2024 · similarity 0.83)
- Neural Collapse for Cross-entropy Class-Imbalanced Learning with Unconstrained ReLU Features Model (ICML 2024 · similarity 0.83)
- I-Con: A Unifying Framework for Representation Learning (ICLR 2025 · similarity 0.83)
- Unlocking Better Closed-Set Alignment Based on Neural Collapse for Open-Set Recognition (AAAI 2025 · 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.