DairySocialTwin: Auditing the video-to-graph evidence chain in cattle social networks
Abstract
Video-derived social graphs can appear precise while inheriting uncertainty from detection, tracking, classification, and aggregation. We present DairySocialTwin, a graph-centered audit framework applied to 203 videos (122 h 23 min) from 15 views at three commercial dairy farms. Curation retained 447 of 839 candidate clips; 400 entered a frozen source-clip index, including 71 legacy clips with unavailable farm and time metadata. Five models were compared on a development split, with an NRI-style model retained as an exploratory trade-off between class balance and dyadic output. In a 90-min manual audit, reconciliation improved switches, fragmentation, and continuous-track duration across three gap definitions, but constrained identity capacity and inherited detector errors limited claims of identity accuracy. Matched manual and reconciled summaries showed coarse agreement and assignment sensitivity. Camera-specific semantics, observation opportunity, and identity-summary-to-video trace-back made these dependencies inspectable. DairySocialTwin supports auditable relational inference without claiming permanent identity, calibrated behavior, or biological diagnosis.
Related
- The Relationship between Personality Traits Expressed in Dairy Cattle and its Effects on Social Enrichment Response: A Scoping Review. Available at SSRN: https://ssrn.com/abstract=6710298
- Genomic evidence of a single domestication origin for modern cattle
- Dairy cattle movement networks govern the spread and persistence of H5N1 B3.13 in the United States
- Identification of differentially expressed lncRNAs and candidate ceRNA networks across Mastitis stages in Sahiwal cattle
- From variants to networks: resolving low SNP overlap in dairy cattle meta-analysis through multi-level functional mapping
BibTeX
@misc{huang2026dairysocialtwin,
title = {{DairySocialTwin: Auditing the video-to-graph evidence chain in cattle social networks}},
author = {Yiwen Huang and Suresh Raja Neethirajan},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.32989},
eprint = {2610.32989},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.32989},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - DairySocialTwin: Auditing the video-to-graph evidence chain in cattle social networks
AU - Yiwen Huang
AU - Suresh Raja Neethirajan
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2610.32989
UR - https://arxiv.org/abs/2610.32989
AB - Video-derived social graphs can appear precise while inheriting uncertainty from detection, tracking, classification, and aggregation. We present DairySocialTwin, a graph-centered audit framework applied to 203 videos (122 h 23 min) from 15 views at three commercial dairy farms. Curation retained 447 of 839 candidate clips; 400 entered a frozen source-clip index, including 71 legacy clips with unavailable farm and time metadata. Five models were compared on a development split, with an NRI-style model retained as an exploratory trade-off between class balance and dyadic output. In a 90-min manual audit, reconciliation improved switches, fragmentation, and continuous-track duration across three gap definitions, but constrained identity capacity and inherited detector errors limited claims of identity accuracy. Matched manual and reconciled summaries showed coarse agreement and assignment sensitivity. Camera-specific semantics, observation opportunity, and identity-summary-to-video trace-back made these dependencies inspectable. DairySocialTwin supports auditable relational inference without claiming permanent identity, calibrated behavior, or biological diagnosis.
ER -