Research map / Label-efficient and robust learning / Recognition and long-tail
Interpretable Part-Whole Hierarchies and Conceptual-Semantic Relationships in Neural Networks
Check 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 |
|---|---|---|---|
| Use Agglomerator instead of Matrix-CapsNet | Image classification models · SmallNorb | error %: 1.40 -> 0.01 | Helped |
| Use Agglomerator instead of E-CapsNet | Image classification models · MNIST | error %: 0.26 -> 0.30 | Hurt |
| Use Agglomerator instead of Capsule VB | Image classification models · FashionMNIST | error %: 5.20 -> 7.43 | Hurt |
| Use Agglomerator instead of ViT-L/16 | Image classification models · CIFAR-10 | error %: 0.85* -> 11.15 | Hurt |
| Use Agglomerator instead of ViT-L/16 | Image classification models · CIFAR-100 | error %: 6.75* -> 40.97 | Hurt |
| Add column layers instead of replacing them with a linear layer (ablation) | Agglomerator · CIFAR-10 | validation loss | Helped |
| Add the contrastive head instead of replacing it with a linear layer (ablation) | Agglomerator · CIFAR-10 | validation loss | Helped |
| Add the linear embedding instead of using its simplified variant (ablation) | Agglomerator · CIFAR-10 | validation loss | Helped |
| Use sinusoidal activations and shared attention (ablation) | Agglomerator · CIFAR-10 | part-whole interpretability | Helped |
| Increase embedding dimension (ablation) | Agglomerator · CIFAR-10 validation | validation loss | Mixed |
Closest papers by meaning
- Neurons to Words: A Novel Method for Automated Neural Network Interpretability and Alignment (AAAI 2025 · similarity 0.86)
- Intrinsic User-Centric Interpretability through Global Mixture of Experts (ICLR 2025 · similarity 0.85)
- Measuring Per-Unit Interpretability at Scale Without Humans (NeurIPS 2024 · similarity 0.85)
- Concept-Guided Interpretability via Neural Chunking (NeurIPS 2025 · similarity 0.84)
- Interpretable Image Classification via Non-parametric Part Prototype Learning (CVPR 2025 · similarity 0.84)
- A Multimodal Automated Interpretability Agent (ICML 2024 · 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.