Research map / Probabilistic and scientific modeling / Variational and generative sampling
FastAMI – a Monte Carlo Approach to the Adjustment for Chance in Clustering Comparison Metrics
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 |
|---|---|---|---|
| Approximate EMI with Monte Carlo sampling instead of exact computation | AMI on large real-world graphs · DBLP | total time [s]: > 7539.9 -> 4.5 | Helped |
| Approximate EMI with Monte Carlo sampling instead of pairwise permutations | AMI on large real-world graphs · DBLP | total time [s]: 361.0 -> 4.5 | Helped |
| Use Monte Carlo EMI estimation rather than pairwise adjustment | AMI on Benchmark Suite for Clustering Algorithms · Benchmark Suite for Clustering Algorithms | Spearman correlation: 0.577 -> 1.000 | Helped |
| Use FastSMI with Monte Carlo variance estimation rather than exact SMI | SMI on Benchmark Suite for Clustering Algorithms · Benchmark Suite for Clustering Algorithms | result returned: 38.3\% -> 99.2\% | Helped |
| Use pairwise permutations to estimate the chance adjustment | AMI · Synthetic clustering comparisons | AMI accuracy | Hurt |
| Use direct Monte Carlo sampling of contingency matrices rather than separate Monte Carlo estimates (ablation) | SMI variance estimation · Synthetic clustering comparisons | runtime | Helped |
| Use separate Monte Carlo estimation of the variance instead of direct contingency-matrix sampling (ablation) | SMI variance estimation · Synthetic clustering comparisons | runtime | Hurt |
| Exploit sparsity by sampling overlaps with the linear mutual-information term absorbed into the distribution | FastAMI EMI estimation · Synthetic clustering comparisons | runtime | Helped |
Closest papers by meaning
- $p$-value Adjustment for Monotonous, Unbiased, and Fast Clustering Comparison (NeurIPS 2023 · similarity 0.89)
- An Evaluative Measure of Clustering Methods Incorporating Hyperparameter Sensitivity (AAAI 2022 · similarity 0.86)
- Generalization Performance of Ensemble Clustering: From Theory to Algorithm (ICML 2025 · similarity 0.85)
- Query-Efficient Correlation Clustering with Noisy Oracle (NeurIPS 2024 · similarity 0.85)
- Expected Probabilistic Hierarchies (NeurIPS 2024 · similarity 0.85)
- Approximately Pareto-optimal Solutions for Bi-Objective k-Clustering (NeurIPS 2024 · similarity 0.85)
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.