Metraipolis › archive › dc_689bcc076dbd
Temporal Island Sparse Autoencoders for Interpreting Clinical Time-Series Models — arXiv:2611.03420
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.03420
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T11:23:16.072Z
- SHA-256 of the archived text
dda8af9a651828468fb9e20056abce5eaf8b33e715dc5d895fc0ee7e7bf71c79- Stable URL of this page
https://metraipolis.com/documento/dc_689bcc076dbd
How to cite
Temporal Island Sparse Autoencoders for Interpreting Clinical Time-Series Models — arXiv:2611.03420 — arXiv (en), https://arxiv.org/abs/2611.03420 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_689bcc076dbd (Metraipolis archive dc_689bcc076dbd, sha256 dda8af9a651828468fb9e20056abce5eaf8b33e715dc5d895fc0ee7e7bf71c79)
BibTeX
@misc{dc689bcc076dbd,
title = {Temporal Island Sparse Autoencoders for Interpreting Clinical Time-Series Models — arXiv:2611.03420},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.03420},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_689bcc076dbd; sha256 dda8af9a651828468fb9e20056abce5eaf8b33e715dc5d895fc0ee7e7bf71c79},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03420. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Deep learning models for electronic health records achieve strong predictive performance but provide little insight into what clinical concepts they represent internally. Post-hoc attribution methods produce per-cell importance scores that lack temporal coherence, cross-variable structure, and reusability across patients. We introduce the Temporal Island Sparse Autoencoder (TI-SAE), which learns a backbone-specific, patient-shared clinical concept dictionary from a frozen predictive model's representations. Each dictionary entry (latent) is defined by learned temporal Gaussian islands (specifying when the concept is active) and a sparse variable-composition vector (specifying which variables are involved). A decoupled selector then identifies which dictionary entries drive each patient's prediction, separating representation learning from attribution. We evaluate TI-SAE across five backbone architectures spanning different representation strategies on MIMIC-III and MIMIC-IV (three prediction tasks each). TI-SAE produces structured clinical concepts that go beyond transient attribution heatmaps while achieving strong performance on standard faithfulness benchmarks. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.03420 — 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 2026-11-03T11:02:37.705Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).