TimeMatch: Unsupervised Cross-region Adaptation by Temporal Shift Estimation

TimeMatch Dataset accompanies the paper 'TimeMatch: Unsupervised Cross-region Adaptation by Temporal Shift Estimation' by Joachim Nyborg of Aarhus University. It includes pre-trained models and data for research on temporal domain adaptation in computer vision. The source code is hosted on GitHub.

Use Cases

• Training unsupervised domain adaptation models based on temporal shift estimation.
• Benchmarking cross-region adaptation algorithms for computer vision tasks.
• Developing methods to handle temporal distribution shifts in image data.

Strengths

• Includes pre-trained models for immediate use.
• Associated with a published academic paper providing methodological context.
• Source code is publicly available on GitHub for reproducibility.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Joachim Nyborg, Aarhus University

Collection Method

Created for the associated research paper.