Authors: Cong Tran, Won-Yong Shin, Andreas Spitz

Published on: June 05, 2021

Impact Score: 8.22

Arxiv code: Arxiv:2106.02926

Summary

• What is new: Proposes IM-META, a novel method for influence maximization in networks with unknown topology using node metadata.
• Why this is important: Finding most influential nodes in a network when its structure is unknown and only a part of it can be explored with limited budget.
• What the research proposes: IM-META addresses this by learning from metadata and nodes’ queries to iteratively construct a network graph and identify influential seed nodes.
• Results: IM-META shows faster network exploration, effectiveness of its modules, superiority over benchmarks, robustness, and scalability across four real-world datasets.

Technical Details

Technological frameworks used: Siamese neural network for learning relationships between metadata and edges.

Models used: nan

Data used: Four real-world datasets for experimental evaluation.

Potential Impact

Social network platforms, marketing agencies, and industries relying on network effects could greatly benefit or need to adapt to these insights.

Want to implement this idea in a business?

We have generated a startup concept here: InfluNet.