Embeddings of Nation-Level Social Networks
2026-04-01T08:52:12Z•e5b6d71d33d12a1f70ae161a93afb14a77acbbb5cfd6c7b24f0078098a4fdebc
LLM-agentsbias-auditcaste-discriminationco-authorship-manipulationdeanonymization-riskdisinformationdual-usedynamic-networksfairness-in-AIgame-theorymultiplex-networksonline-network-designprivacyred-teamingsocial-network-embeddingssocial-simulationsynthetic-social-observatorytemporal-embeddingstransport-optimization
What happened
This collection of arXiv abstracts covers nation-scale social network embeddings, a real-time LLM-driven social-agent simulator (BotVerse), a game-theoretic model of co-authorship dynamics, an LLM audit of caste bias in matchmaking, and online design methods for dynamic networks. Technical contributions include layer-sensitive and temporal alignment strategies for multiplex/temporal node embeddings (nation-level privacy risk), an event-driven orchestration and simulation engine for deploying large populations of LLM agents grounded in live content streams (enables scalable red-teaming but also
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- arxiv_cs_si
- Record identifier
- e5b6d71d33d12a1f70ae161a93afb14a77acbbb5cfd6c7b24f0078098a4fdebc
- Enrichment time
- 2026-04-01T08:52:12Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.