Metraipolis › archive › dc_d6cfd341f12d
TAP-Path: Task-Adaptive Structural and Token Pruning for Efficient and Trustworthy Pathology Foundation Models — arXiv:2609.04071
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.04071
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T01:12:15.770Z
- SHA-256 of the archived text
1f3fa4b019a6d5ba4e0983188a075ac84b6bf8b7161491093667ab2cf502aea6- Stable URL of this page
https://metraipolis.com/documento/dc_d6cfd341f12d
How to cite
TAP-Path: Task-Adaptive Structural and Token Pruning for Efficient and Trustworthy Pathology Foundation Models — arXiv:2609.04071 — arXiv (en), https://arxiv.org/abs/2609.04071 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d6cfd341f12d (Metraipolis archive dc_d6cfd341f12d, sha256 1f3fa4b019a6d5ba4e0983188a075ac84b6bf8b7161491093667ab2cf502aea6)
BibTeX
@misc{dcd6cfd341f12d,
title = {TAP-Path: Task-Adaptive Structural and Token Pruning for Efficient and Trustworthy Pathology Foundation Models — arXiv:2609.04071},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.04071},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d6cfd341f12d; sha256 1f3fa4b019a6d5ba4e0983188a075ac84b6bf8b7161491093667ab2cf502aea6},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.04071. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Pathology foundation models improve transferable representation learning for histopathology, but recent gains often rely on encoders with hundreds of millions of parameters and high inference cost. We propose TAP-Path, a task-adaptive compression framework that directly restructures a pretrained Virchow2 encoder rather than distilling it into a separate student. TAP-Path combines validation-driven transformer-block selection, physical removal of redundant blocks, input-adaptive patch-token pruning, multi-depth feature recovery, and a lightweight gated task head. The final model retains 24 of 32 transformer blocks and 70% of patch tokens after pruning, reducing encoder parameters by 24.96% (631.24M to 473.70M) and analytical encoder compute by 35.20% (340.13G to 220.40G FLOPs). Across three task-head optimization seeds, TAP-Path achieved $87.98 \pm 0.067%$ test accuracy, $81.26 \pm 0.49%$ balanced accuracy, and $82.38 \pm 0.48%$ macro-F1 on a 32-class histopathology benchmark, compared with 86.89% for full Virchow2 and 87.67% for UNI2-h. TAP-Path achieved a Brier score of $0.1800 \pm 0.0005$ and failure-detection AUROC of $0.9047 \pm 0.0060$. A validation-only rare-aware objective improved rare-class balanced accuracy in a secondary operating analysis. Frozen external evaluation on 433 CPTAC samples yielded $91.22 \pm 0.83%$ accuracy and $91.10 \pm 0.81%$ balanced accuracy. These results show that task-adaptive structural and token sparsification can improve the accuracy-efficiency trade-off of large pathology foundation models while preserving reliability under internal and external evaluation. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-04 in cs.AI). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.04071 — 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 2026-10-17T00:28:03.029Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-04?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).