Research map / Learning theory and algorithms / Prediction and decision losses
Diffusion Source Identification on Networks with Statistical Confidence
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 an ADiT confidence set instead of a Euclidean confidence set | Diffusion source identification · Random 4-regular tree, 90% confidence | average set size: 79.1 -> 60.4 | Helped |
| Use an ADiT confidence set instead of a Euclidean confidence set | Diffusion source identification · Preferential attachment, 90% confidence | average set size: 86.2 -> 70.5 | Helped |
| Use an ADiT confidence set instead of a Euclidean confidence set | Diffusion source identification · Small-world network, 90% confidence | average set size: 16.0 -> 17.7 | Hurt |
| Use an ADiT confidence set instead of a Euclidean confidence set | Diffusion source identification · Random 4-regular tree, 80% confidence | average set size: 68.6 -> 50.0 | Helped |
| Use an ADiT confidence set instead of a Euclidean confidence set | Diffusion source identification · Preferential attachment, 80% confidence | average set size: 71.9 -> 57.5 | Helped |
| Use an ADiT confidence set instead of a Euclidean confidence set | Diffusion source identification · Small-world network, 80% confidence | average set size: 10.5 -> 11.3 | Hurt |
| Use confidence sets rather than a single-point source estimate | Diffusion source identification · Random network models | source coverage: ADiT-90%: 93%, 89%, 91%; single-point ADiT: 0, 0.02, 0.16 | Helped |
| Use both importance sampling and isomorphism pooling instead of vanilla Monte Carlo | Confidence-set construction · Random 4-regular tree | sequential running time (sec.): 898.3 -> 411.9 | Helped |
| Use both importance sampling and isomorphism pooling instead of vanilla Monte Carlo | Confidence-set construction · Preferential attachment | sequential running time (sec.): 1060.5 -> 430.9 | Helped |
| Use both importance sampling and isomorphism pooling instead of vanilla Monte Carlo | Confidence-set construction · Small-world network | sequential running time (sec.): 1052.9 -> 1057.7 | No clear effect |
| Use inverse-time ADiT weighting instead of constant-weight Euclidean discrepancy | Single-point source estimation · Preferential attachment | correct rate: 0.04 -> 0.02 | Hurt |
| Use inverse-time ADiT weighting instead of constant-weight Euclidean discrepancy | Single-point source estimation · Small-world network | correct rate: 0.20 -> 0.16 | Hurt |
Closest papers by meaning
- Conformal Prediction for Multi-Source Detection on a Network (AAAI 2026 · similarity 0.88)
- Learning Diffusions under Uncertainty (AAAI 2024 · similarity 0.88)
- Diffusion Network Inference from Partial Observations (AAAI 2021 · similarity 0.87)
- Learning Diffusions without Timestamps (AAAI 2019 · similarity 0.86)
- Diffusion Model for Graph Inverse Problems: Towards Effective Source Localization on Complex Networks (NeurIPS 2023 · similarity 0.85)
- Diffusion Attribution Score: Evaluating Training Data Influence in Diffusion Models (ICLR 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.