CUB Inc10: Representative Failure Cases for Class-Incremental Learning

A 5.5 KB Excel file containing representative failure cases from the CUB Inc10 benchmark for Class-Incremental Learning (CIL). The dataset was authored by Jianye Gu and last updated on May 7, 2026, under a CC-BY-4.0 license. It likely contains examples of misclassifications to illustrate the challenge of catastrophic forgetting in CIL models.

Use Cases

• Analyzing model failure patterns based on the described benchmark for Class-Incremental Learning.
• Evaluating the resistance to catastrophic forgetting based on the described performance metrics.
• Benchmarking lightweight adapter methods like DGASA based on the described experimental framework.

Strengths

• Dataset is small (5.5 KB), facilitating quick download and inspection.
• Specifically focuses on failure cases, providing targeted diagnostic data.
• Released under a permissive CC-BY-4.0 license for broad reuse.

Limitations

• Row count and column-level documentation are absent; field semantics must be inferred after download.
• The dataset's scope is extremely limited, containing only failure cases from one benchmark.

Provenance

Source

figshare

Collection Method

Likely extracted from experimental results of a Class-Incremental Learning study.

Time Range

null

Freshness

Last updated 2026-05-07 17:41:50; freshness should be verified.

Geography

null