Ethical Considerations for Responsible Data Curation

Contextualization
DOI: 10.48550/arxiv.2302.03629 Publication Date: 2023-01-01
ABSTRACT
Human-centric computer vision (HCCV) data curation practices often neglect privacy and bias concerns, leading to dataset retractions unfair models. HCCV datasets constructed through nonconsensual web scraping lack crucial metadata for comprehensive fairness robustness evaluations. Current remedies are post hoc, persuasive justification adoption, or fail provide proper contextualization appropriate application. Our research focuses on proactive, domain-specific recommendations, covering purpose, consent, diversity, curating evaluation datasets, addressing concerns. We adopt an ante hoc reflective perspective, drawing from current practices, guidelines, withdrawals, audits, inform our considerations recommendations.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES ()
CITATIONS ()
EXTERNAL LINKS
PlumX Metrics
RECOMMENDATIONS
FAIR ASSESSMENT
Coming soon ....
JUPYTER LAB
Coming soon ....