Calibration Analysis on CUB Inc10: Class-Incremental Learning Evaluation

5.5 KB Excel file containing calibration analysis results for the CUB Inc10 benchmark, authored by Jianye Gu and last updated in May 2026. The data supports the evaluation of a proposed method, DGASA, for mitigating catastrophic forgetting in Class-Incremental Learning tasks.

Use Cases

• Benchmarking new CIL methods against DGASA based on the described calibration analysis.
• Analyzing prototype drift and subspace alignment in continual learning based on the described feature subspaces.
• Evaluating computational efficiency and generalization in CIL models based on the described experimental results.

Strengths

• File size is precisely 5.5 KB, indicating a compact and focused dataset.
• Released under the permissive CC-BY-4.0 license for broad reuse.
• Associated with a specific, described machine learning method (DGASA) for context.

Limitations

• Row count and column definitions are unknown, requiring manual inspection after download.
• Description metadata is limited; actual data quality and structure require verification.
• The dataset's small scale (5.5 KB) suggests it contains summary results, not raw training data.

Provenance

Source

figshare, authored by Jianye Gu.

Collection Method

Likely generated from experiments evaluating the DGASA method on the CUB Inc10 benchmark.

Time Range

null

Freshness

Last updated 2026-05-07 17:41:51; freshness should be verified.

Geography

null