Metraipolis › archive › dc_ecd079d3188b

Hippocampal consolidation without replay in fully fine-tuned networks — arXiv:2611.01952

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.01952
License (as archived)
CC0-1.0
Archived
2026-11-03T10:52:47.119Z
SHA-256 of the archived text
6dc82b483c9ce7d5a7ce2ca40be0bfda991c0b98b783ec372fd0e9652e415d57
Stable URL of this page
https://metraipolis.com/documento/dc_ecd079d3188b

How to cite

Hippocampal consolidation without replay in fully fine-tuned networks — arXiv:2611.01952 — arXiv (en), https://arxiv.org/abs/2611.01952 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ecd079d3188b (Metraipolis archive dc_ecd079d3188b, sha256 6dc82b483c9ce7d5a7ce2ca40be0bfda991c0b98b783ec372fd0e9652e415d57)
BibTeX
@misc{dcecd079d3188b,
  title = {Hippocampal consolidation without replay in fully fine-tuned networks — arXiv:2611.01952},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.01952},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ecd079d3188b; sha256 6dc82b483c9ce7d5a7ce2ca40be0bfda991c0b98b783ec372fd0e9652e415d57},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.01952. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Networks that keep learning overwrite what they know. Prevailing remedies replay stored data, distil from old models or freeze the backbone. The brain suggests a different division of labour: the hippocampus indexes where experience lies in cortex while the cortex keeps learning. We show that this division lets a fully trainable network keep learning without cross-task replay, a teacher or inference-time memory. After each task, a training-only hippocampal memory indexes the input geometry of every group of synapses sharing an input, residual energy included, and limits each synapse's movement per direction before and after the optimizer. This class-free geometry carries most of the protection; with conflict detection on top, the method cuts absorption of established classes by near-duplicates several-fold against every compared method. To our knowledge, this is the first demonstration of one principle keeping whole networks learning across four architecture families and models up to 8 billion parameters. Methods that freeze the backbone hold up only as far as its representation already separates the classes; here the whole network keeps learning to tell near-duplicates apart, leading the frozen-backbone baseline by 17.5 points on TACRED. Without the memory, much old knowledge stays decodable but beyond the network's reach.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.01952 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-03T10:49:38.325Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).