Research map / Probabilistic and scientific modeling / Variational and generative sampling
Co-Occurrence Estimation from Aggregated Data with Auxiliary Information
Tomoharu Iwata, Naoki Marumo
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 |
|---|---|---|---|
| add auxiliary item features (EB+CGM to Proposed) (ablation) | Proposed method · Sushi U*=20 | absolute error: 34% reduction | Improved |
| add auxiliary item features (EB+CGM to Proposed) (ablation) | Proposed method · Movielens U*=20 | absolute error: 8% reduction | Improved |
| use aggregated marginal counts (Proposed vs ML, EB, EB+A) | Proposed method · Sushi and Movielens | absolute error | Improved |
| increase number of users with records U* | Proposed method · Sushi and Movielens | absolute error | Improved |
| increase number of items I | Proposed method · Sushi, U*=20 | absolute error | Improved |
| set lambda weight to 0 (EB+A) or 1 (CGM+A) rather than intermediate (ablation) | Proposed method · Sushi | absolute error: best lambda = 0.4 | Improved |
| set lambda weight to 0.1 rather than 0 or 1 (ablation) | Proposed method · Movielens | absolute error: best lambda = 0.1 | Improved |
| use Proposed instead of IND | co-occurrence estimation · Sushi pairs, U*=20 | number of best pairs: 15 -> 30 | Improved |
| use Proposed instead of CGM+A | co-occurrence estimation · Movielens U*=20 I=200 | rate of item pairs best: 0.211 -> 0.242 | Improved |
| use Proposed instead of IND (independence assumption) | co-occurrence estimation · shrimp-squid pair | absolute error: 0.035 -> 0.165 | Hurt |
| use Proposed instead of IND | co-occurrence estimation · shrimp-sea eel pair | absolute error: 0.129 -> 0.027 | Improved |
| use Proposed on larger data | co-occurrence estimation · Movielens-1M | absolute error: 0.0169 -> 0.0147 | Improved |
Closest papers by meaning
- Exploiting Negative Samples: A Catalyst for Cohort Discovery in Healthcare Analytics (ICML 2024 · similarity 0.83)
- Sparse Parameter Recovery from Aggregated Data (ICML 2016 · similarity 0.83)
- ACES: Automatic Cohort Extraction System for Event-Stream Datasets (ICLR 2025 · similarity 0.83)
- NeuralCohort: Cohort-aware Neural Representation Learning for Healthcare Analytics (ICML 2025 · similarity 0.82)
- InfoNet: Neural Estimation of Mutual Information without Test-Time Optimization (ICML 2024 · similarity 0.81)
- Approximating mutual information of high-dimensional variables using learned representations (NeurIPS 2024 · 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.
