Research map / Learning theory and algorithms / Submodular and game algorithms
Learning to Elect
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 |
|---|---|---|---|
| Use a permutation-invariant network rather than an MLP | Mimicking classical voting rules · Borda, Copeland, Maximin, and Kemeny (synthetic uniform) | accuracy: 0.95 -> 0.99; 0.81 -> 0.83; 0.75 -> 0.80; 0.76 -> 0.94 | Helped |
| Use a permutation-invariant network rather than an MLP | Mimicking Plurality · Plurality (synthetic uniform) | accuracy: 1.0 -> 1.0 | No clear effect |
| Train a Set Transformer on synthetic elections with fewer than 100 voters | Borda winner prediction · Borda (100-199 voters) | accuracy: 0.99; 0.99 | Helped |
| Train on synthetic elections and test zero-shot on real election datasets | Learned voting rules · Netflix, Sushi, and Mechanical Turk Puzzle | accuracy | Helped |
| Learn a voting rule from an oracle for the target social welfare | Neural voting rules · Utilitarian and egalitarian welfare across uniform, polarized, and indecisive distributions | accuracy | Helped |
| Use a neural voting rule rather than the strongest listed classical rule | Utilitarian welfare prediction · Uniform, polarized, and indecisive distributions | accuracy: 0.56 -> 0.69; 0.50 -> 0.68; 0.61 -> 0.70 | Helped |
| Use a neural voting rule rather than the strongest listed classical rule | Egalitarian welfare prediction · Uniform, polarized, and indecisive distributions | accuracy: 0.58 -> 0.66; 0.51 -> 0.68; 0.53 -> 0.57 | Helped |
| Use the learned network rather than the theoretical utilitarian-optimal rule | Utilitarian welfare prediction · Uniform, polarized, and indecisive distributions | accuracy: 0.65 -> 0.69; 0.65 -> 0.68; 0.67 -> 0.70 | Helped |
| Replace BatchNorm with LayerNorm in GIN (ablation) | GIN · Training stability and performance | performance | No clear effect |
| Use Lookahead optimization (ablation) | DeepSets · Training | training effectiveness | Helped |
| Use Lookahead optimization (ablation) | GIN and Set Transformer · Training | training effectiveness | No clear effect |
| Use one-hot candidate IDs instead of integer candidate IDs (ablation) | Permutation-invariant voting networks · Voting rule learning | performance | Helped |
Closest papers by meaning
- DeepVoting: Learning and Improving Voting Rules with Fine-Tuning (NeurIPS 2026 · similarity 0.88)
- Learning to Manipulate Under Limited Information (AAAI 2025 · similarity 0.85)
- Discerning Decision-Making Process of Deep Neural Networks with Hierarchical Voting Transformation (NeurIPS 2021 · similarity 0.85)
- Weighted Voting Via No-Regret Learning (AAAI 2018 · similarity 0.84)
- Explainable Voting (NeurIPS 2020 · similarity 0.84)
- C-Voting: Confidence-Based Test-Time Voting without Explicit Energy Functions (ICLR 2026 · similarity 0.81)
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.