Research map / Label-efficient and robust learning / Recognition and long-tail
Shape Constraints for Set Functions
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 trapezoid constraints on star rating and word count (ablation) | DLN · Kaggle puzzles test | MSE: 8330 -> 7587 | Helped |
| Use DLN set function instead of Deep Sets | SFE set function · Kaggle puzzles test | MSE: 8323 -> 8330 | No clear effect |
| Use DLN instead of Deep Sets on Kickstarter | SFE set function · Kickstarter test | accuracy (%): 65.7 -> 65.8 | Helped |
| Use SFE with learned set function instead of DNN on Kickstarter | SFE + DLN · Kickstarter test | accuracy (%): 64.9 -> 65.8 | Helped |
| Use learned set function instead of mean token estimate on Kickstarter | SFE · Kickstarter test | accuracy (%): 64.3 -> 65.8 | Helped |
| Add trapezoid constraints on both ngram order and frequency (ablation) | DLN · Kickstarter test | accuracy (%): 65.8 -> 65.8 | No clear effect |
| Add trapezoid on ngram frequency (ablation) | DLN · Kickstarter test | accuracy (%): 65.8 -> 65.9 | Helped |
| Add trapezoid constraints on frequency (ablation) | DLN · Recipes test | precision@3 (%): 88.8 -> 88.3 | Hurt |
| Add trapezoid constraints on both (ablation) | DLN · Recipes test | precision@3 (%): 88.8 -> 88.0 | Hurt |
| Use DNN instead of DLN | DLN · Recipes test | precision@3 (%): 88.8 -> 88.9 | Helped |
| Use DLN instead of Deep Sets | SFE set function · Recipes test | precision@3 (%): 87.7 -> 88.8 | Helped |
| Use learned DLN instead of mean SFE token estimate | SFE · Recipes test | precision@3 (%): 87.5 -> 88.8 | Helped |
Closest papers by meaning
- Neural Set Function Extensions: Learning with Discrete Functions in High Dimensions (NeurIPS 2022 · similarity 0.87)
- Deep Functional Dictionaries: Learning Consistent Semantic Structures on 3D Models from Functions (NeurIPS 2018 · similarity 0.86)
- Deep Sets (NeurIPS 2017 · similarity 0.86)
- Polynomial Width is Sufficient for Set Representation with High-dimensional Features (ICLR 2024 · similarity 0.84)
- Empowering Decision Trees via Shape Function Branching (NeurIPS 2025 · similarity 0.83)
- Monotone and Separable Set Functions: Characterizations and Neural Models (NeurIPS 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.