research
The Replay Buffer Refreshed Its Old Examples
Uniform Herding reselects bounded class exemplars in the model's current representation instead of freezing yesterday's geometry.
Summary
Uniform Herding reselects bounded class exemplars in the model's current representation instead of freezing yesterday's geometry.
On a ten-task CIFAR-100 protocol with a 2,000-example active budget, the method reports 44.00 percent final average accuracy and 17.22 percent forgetting, compared with 42.33 percent and 24.87 percent for iCaRL. Removing distillation increased forgetting, while changes to the active budget mattered more than the retrieval budget. Because the end-to-end comparison changes more than exemplar refresh alone, the paper explicitly limits its causal claim to the tested protocol.
Why it matters
Uniform Herding reselects bounded class exemplars in the model's current representation instead of freezing yesterday's geometry.
Limits and context
No additional limitation was separately recorded.
Key claims
Uniform Herding reselects bounded class exemplars in the model's current representation instead of freezing yesterday's geometry.
Evidence: source-2026-08-14-010
Sources
- arXiv preprint 2608.13061arXiv · primary research
Corrections
No corrections have been recorded for this story.