Metraipolis › archive › dc_c95ac5b071bf

Drift-adaptive Representation Preservation for Domain-Incremental Learning — arXiv:2611.06093

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.06093
License (as archived)
CC0-1.0
Archived
2026-11-05T18:17:44.253Z
SHA-256 of the archived text
d0ec50eae0122285103d44811aacd990bbd60638decee96df3215759f2add225
Stable URL of this page
https://metraipolis.com/documento/dc_c95ac5b071bf

How to cite

Drift-adaptive Representation Preservation for Domain-Incremental Learning — arXiv:2611.06093 — arXiv (en), https://arxiv.org/abs/2611.06093 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c95ac5b071bf (Metraipolis archive dc_c95ac5b071bf, sha256 d0ec50eae0122285103d44811aacd990bbd60638decee96df3215759f2add225)
BibTeX
@misc{dcc95ac5b071bf,
  title = {Drift-adaptive Representation Preservation for Domain-Incremental Learning — arXiv:2611.06093},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.06093},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c95ac5b071bf; sha256 d0ec50eae0122285103d44811aacd990bbd60638decee96df3215759f2add225},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.06093. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Pretrained vision models provide strong general-purpose representations, but in Domain-Incremental Learning (DIL), where the input distribution changes sequentially, adapting to new domains often degrades performance on previously learned ones. We observe that forgetting is not uniform across domains and classes and is not explained by training order alone. This motivates us to track each past domain-class concept individually. We use representation drift, defined as the deviation of a past concept's current representation from a fixed anchor, as a measurable surrogate for forgetting risk. The measured drift then serves two complementary roles. A consolidation loss assigns stronger preservation to high-drift pairs, while prototype construction down-weights highly drifted domains to reduce their influence on the class reference. Preservation is thus allocated by measured drift rather than assumed importance. To obtain this signal during training without storing past raw data, we repurpose a Joint-Embedding Predictive Architecture (JEPA). For each domain-class pair, we store only a compact context summary in the backbone feature space and rerun it through the learned context-to-target relation to measure its drift from the fixed anchor. Across DomainNet, Office-Home, CORe50, and CDDB-Hard, our method sustains a strong accuracy-forgetting balance on every benchmark, using substantially less training memory than accuracy-competitive methods and the smallest storage among compared methods.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.06093 — 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 2026-11-05T18:12:49.046Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).