Searcharxiv⌕ Search

arXiv subjects

Kasra Ghorbani

Publications and source records attributed to Kasra Ghorbani.

2 recordsLinked to original sources

Locating Debunkers for Rumor Mitigation in Directed Networks

Rumors can spread widely through social networks, making the selection of a limited number of debunkers to disseminate corrective information important. Existing centrality-based methods only consider the individual influence of a node on spreading processes. Those individually influential nodes often have heavily overlapping influence ranges, resulting in a limited number of users they can cover. To address this problem, we propose the so-called Hybrid Marginal Influence (HMI) framework, which combines node centrality with additional audience coverage provided by each candidate and adaptively balances these two factors through a greedy selection procedure. We further develop the Debunker-Augmented SIR (DA-SIR) model for the propagation of rumors and corrective information. Experiments on ten real-world directed networks show that HMI generally selects debunkers with broader audience coverage than conventional centrality-based methods, accompanied by a lower final cumulative fraction of users who repost the rumor.

physics.soc-ph↗

Bison: Cross-Dataset Learning for Unseen-Compound Perturbation Prediction

Predicting transcriptional responses to unseen compounds is limited by fragmented chemical coverage and heterogeneous experimental platforms and gene panels. To assess molecular generalization across these settings, we build on Chem-PerturBridge to benchmark eight datasets with 16,771 compounds, withholding test compounds from every training dataset. This comparison reveals that high overall response agreement can coexist with weak prediction of drug-specific differences, despite reproducible signals across repeated measurements. To exploit complementary chemical supervision while targeting these differences, we introduce Bison: a shared gene representation connects native panels, while two discrete diffusion models compose context-dependent responses with molecular deviations learned through matched drug-contrast supervision. A single Bison model jointly trained across all eight datasets achieves the highest mean overall-response and drug-contrast Pearson correlations on the full benchmark in comparison with 11 methods trained independently per dataset. Compared with dataset-specific training of the same architecture, joint training increases mean drug-contrast correlation by 27.4\%, with gains across all eight datasets and improvements in overall response prediction. These results demonstrate how matched drug contrasts turn complementary screens into shared molecular supervision for unseen-drug response prediction while preserving native gene measurements.

cs.LG↗