Research map / Probabilistic and scientific modeling / Variational and generative sampling
Cooperative neural networks (CoNN): Exploiting prior independence structure for improved classification
Harsh Shrivastava, Eugene Bart, Bob Price, Hanjun Dai, Bo Dai, Srinivas Aluru
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 CoNN-sLDA instead of the strongest reported baseline, DUI-sLDA (L=5) | MultiSent classification · MultiSent | AUC: 91.4$\pm$ 0.27 -> 93.3$\pm$ 0.13 | Improved |
| Use CoNN-sLDA instead of the strongest reported baseline, DUI-sLDA | 20 Newsgroups classification · 20 Newsgroups | accuracy: 83.5$\pm$ 0.22 -> 83.4 $\pm$ 0.18 | No clear effect |
| Add cost-sensitive class weighting (ablation) | CoNN-sLDA · 20 Newsgroups | accuracy: 83.4 $\pm$ 0.18 -> 83.7$\pm$ 0.13 | Improved |
| Add cost-sensitive class weighting (ablation) | CoNN-sLDA · MultiSent | AUC: 93.3$\pm$ 0.13 -> 93.4$\pm$ 0.13 | Improved |
| Increase the number of update iterations (ablation) | CoNN-sLDA · MultiSent | AUC | Hurt |
| Unroll the update algorithm beyond 5 iterations (ablation) | CoNN-sLDA · MultiSent | AUC | No clear effect |
| Use deeper embedding networks (ablation) | CoNN-sLDA · MultiSent | AUC | Hurt |
| Use two fully connected layers for the word-topic embedding (ablation) | CoNN-sLDA · MultiSent | AUC | Improved |
| Increase Hilbert-space dimension beyond 20 (ablation) | CoNN-sLDA · MultiSent | AUC | Hurt |
Closest papers by meaning
- DNNs as Layers of Cooperating Classifiers (AAAI 2020 · similarity 0.85)
- Cooperative Sheaf Neural Networks (ICLR 2026 · similarity 0.85)
- Cooperative Graph Neural Networks (ICML 2024 · similarity 0.83)
- Adversarial Cooperative Rationalization: The Risk of Spurious Correlations in Even Clean Datasets (ICML 2025 · similarity 0.83)
- Graph Metanetworks for Processing Diverse Neural Architectures (ICLR 2024 · similarity 0.83)
- Unified Graph Neural Networks Pre-training for Multi-domain Graphs (AAAI 2025 · similarity 0.82)
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.
