Sensitivity of Diffusion Dynamics to Network Uncertainty
Authors: A. Adiga, C. J. Kuhlman, H. S. Mortveit, A. K. S. Vullikanti
JAIR 2014 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | (3) We study these sensitivity questions using extensive simulations on diverse real world networks and find that our theoretical predictions for both models match the observations quite closely. (4) Experimentally, the transient behavior, i.e., the time series of the number of infections, in both models appears to be more sensitive to network perturbations. 5. Experimental Results |
| Researcher Affiliation | Academia | Abhijin Adiga EMAIL Chris J. Kuhlman EMAIL Network Dynamics and Simulation Science Laboratory, Virginia Bioinformatics Institute, Virginia Tech, VA 24061 Henning S. Mortveit EMAIL Network Dynamics and Simulation Science Laboratory, Virginia Bioinformatics Institute, and Department of Mathematics, Virginia Tech, VA 24061 Anil Kumar S. Vullikanti EMAIL Network Dynamics and Simulation Science Laboratory, Virginia Bioinformatics Institute, and Department of Computer Science, Virginia Tech, VA 24061 |
| Pseudocode | No | The paper describes the Independent Cascade (IC) and Linear Threshold (LT) models in text, but no structured pseudocode or algorithm blocks are provided. |
| Open Source Code | No | The paper does not provide explicit statements or links to source code for the described methodology. A technical report is mentioned for the full version, but this does not indicate code release: "Adiga, A., Kuhlman, C., Mortveit, H. S., & Vullikanti, A. K. S. (2014). Sensitivity of diffusion dynamics to network uncertainty. Technical report, available at http://ndssl.vbi.vt.edu/supplementary-info/vskumar/sensitivity-jair.pdf." |
| Open Datasets | Yes | We study the sensitivity to edge perturbations on twenty diverse real-world networks (Leskovec, 2011) with varying degrees of perturbation and other factors for both IC and LT models. |
| Dataset Splits | No | For each diffusion instance, the seed set was constructed by sampling the vertex set uniformly with probability s, and every node was assigned a threshold chosen uniformly at random in the interval [0, 1]. This describes the sampling method for initial seed nodes but not explicit dataset splits (e.g., train/test/validation) for the networks used. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory) used to run the experiments or simulations. |
| Software Dependencies | No | The paper does not specify any software dependencies with version numbers (e.g., programming languages, libraries, frameworks, or solvers). |
| Experiment Setup | Yes | Each network G in Table 1 was perturbed with values of ϵ ranging from 0 to 100, where ϵ = 0 corresponds to the unperturbed network. For each ϵ, we generated ten graph instances G = G + R or G R. Here, R may be Ru or Rd, depending on whether the perturbation is a uniform edge approach or a degree-assortative approach. For each graph instance, we performed a simulation run, which consists of 100 separate diffusion instances. |