Dataset Condensation Atlas

Application · Meta-learning through unrolled training

Distilled Replay: Overcoming Forgetting through Synthetic Samples

Andrea Rosasco, Antonio Carta, Andrea Cossu, Vincenzo Lomonaco, Davide Bacciu

IJCAI 2021 Workshop · first public 2021-03-29 · arXiv 2103.15851

paper ↗code ↗catalogued✓ abstract read

In one paragraph

Builds a rehearsal buffer for continual learning by distilling the full dataset into one highly informative synthetic pattern per class, rather than storing real exemplars, keeping the buffer extremely small. Reports effectiveness against popular replay-based continual learning baselines on four benchmarks.

Where it sits

Abstract (verbatim from arXiv)

Replay strategies are Continual Learning techniques which mitigate catastrophic forgetting by keeping a buffer of patterns from previous experiences, which are interleaved with new data during training. The amount of patterns stored in the buffer is a critical parameter which largely influences the final performance and the memory footprint of the approach. This work introduces Distilled Replay, a novel replay strategy for Continual Learning which is able to mitigate forgetting by keeping a very small buffer (1 pattern per class) of highly informative samples. Distilled Replay builds the buffer through a distillation process which compresses a large dataset into a tiny set of informative examples. We show the effectiveness of our Distilled Replay against popular replay-based strategies on four Continual Learning benchmarks.

BibTeX (generated; prefer the venue's official entry)
@article{rosasco2021distilled,
  title   = {Distilled Replay: Overcoming Forgetting through Synthetic Samples},
  author  = {Andrea Rosasco and Antonio Carta and Andrea Cossu and Vincenzo Lomonaco and Davide Bacciu},
  journal = {IJCAI 2021 Workshop},
  year    = {2021}
}

Nearby in Meta-learning through unrolled training

2024-10

Offline Behavior Distillation

Shiye Lei, Sen Zhang, Dacheng Tao · NeurIPS 2024notableOther datapaper ↗code ↗

2023-11

RaT-BPTT — Embarassingly Simple Dataset Distillation

Yunzhen Feng, Ramakrishna Vedantam, Julia Kempe · ICLR 2024notablepaper ↗code ↗

2019-12

GTN — Generative Teaching Networks: Accelerating Neural Architecture Search by Learning to Generate Synthetic Training Data

Felipe Petroski Such, Aditya Rawal, Joel Lehman et al. · ICML 2020notablepaper ↗code ↗

2018-11

DD — Dataset Distillation

Tongzhou Wang, Jun-Yan Zhu, Antonio Torralba et al. · arXiv 2018landmarkpaper ↗code ↗