Metraipolis › archive › dc_d42283245b98
Operational Abstractions of Neural Network Concepts via Topological Representations — arXiv:2610.05017
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.05017
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T05:26:37.631Z
- SHA-256 of the archived text
0bec2fa9cad862fae36c4f15521f60a031b93a4d1f2550bb0edab4048ae75d35- Stable URL of this page
https://metraipolis.com/documento/dc_d42283245b98
How to cite
Operational Abstractions of Neural Network Concepts via Topological Representations — arXiv:2610.05017 — arXiv, https://arxiv.org/abs/2610.05017 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d42283245b98 (Metraipolis archive dc_d42283245b98, sha256 0bec2fa9cad862fae36c4f15521f60a031b93a4d1f2550bb0edab4048ae75d35)
BibTeX
@misc{dcd42283245b98,
title = {Operational Abstractions of Neural Network Concepts via Topological Representations — arXiv:2610.05017},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.05017},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d42283245b98; sha256 0bec2fa9cad862fae36c4f15521f60a031b93a4d1f2550bb0edab4048ae75d35},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.05017. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Concept-based methods provide a semantic level for interpreting and manipulating learned representations, but existing editing approaches are typically specialized to particular interventions and do not provide a common and editable representation of concept organization. To achieve this, we introduce Topological Concept Representations (TCR), a post-hoc operational abstraction that jointly characterizes the concepts encoded in a learned representation and their relationships. TCR constructs an intermediate concept space from concept recoverability and interaction scores, and compactly encodes its organization through topology. Interventions are expressed through modifications of this abstraction that are propagated back to the underlying learned representation. This allows different concept-level operations to share the same optimization framework and separates the desired concept organization from the mechanism to achieve it. We establish stability and reparameterization-invariance properties of TCR and its connections to existing concept-editing formulations. We use TCR to disentangle concepts as a preprocessing step for existing erasure methods, improving worst-group accuracy by 21.89 on average at comparable concept leakage. We further use TCR to transfer concepts from teacher to student models, improving concept recoverability by up to 5.54 while also improving or maintaining competitive test top-1 accuracy. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.05017 — 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 2026-10-06T05:24:05.955Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).