Fractional physics-informed neural networks to model social media virality
Source check: Source checked. Not independent scientific confirmation.
Finding: A segmented fractional physics-informed neural network modeled engagement trajectories for 348 X posts. It reconstructed 64% of held-out trajectories within the stated error threshold, versus 44% for the strongest of nine classical baselines. The result is limited to the sampled posts and does not establish causal social-media dynamics.
Technology: Fractional physics-informed neural network with symbolic regression
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A fractional PINN learned temporal memory and nonlinear engagement dynamics; symbolic regression translated learned dynamics into equations.
- Test and result
- On 348 X posts, a four-segment teacher-forced model reconstructed 64% of held-out trajectories within the reported mean-squared-error threshold, versus 44% for the strongest of nine classical baselines. Synthetic experiments also tested recovery under clean and noisy observations.
- Limitation / still unknown
- The single-window model performed poorly on held-out trajectories; results concern sampled posts and do not establish causal mechanisms or general social-media dynamics.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://arxiv.org/abs/2611.07215v1 · https://arxiv.org/html/2611.07215v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.