Metraipolis › archive › dc_887b372bd703

Attributing HOW, Not Just WHICH: Counterfactual Response Trajectories for Diffusion Models — arXiv:2610.11238

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.11238
License (as archived)
CC0-1.0
Archived
2026-10-09T06:30:25.089Z
SHA-256 of the archived text
b2d9cfca2f85d91fad2a097513952b474cd03e16d35a5720f1e4bf3daab90cd9
Stable URL of this page
https://metraipolis.com/documento/dc_887b372bd703

How to cite

Attributing HOW, Not Just WHICH: Counterfactual Response Trajectories for Diffusion Models — arXiv:2610.11238 — arXiv (en), https://arxiv.org/abs/2610.11238 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_887b372bd703 (Metraipolis archive dc_887b372bd703, sha256 b2d9cfca2f85d91fad2a097513952b474cd03e16d35a5720f1e4bf3daab90cd9)
BibTeX
@misc{dc887b372bd703,
  title = {Attributing HOW, Not Just WHICH: Counterfactual Response Trajectories for Diffusion Models — arXiv:2610.11238},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.11238},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_887b372bd703; sha256 b2d9cfca2f85d91fad2a097513952b474cd03e16d35a5720f1e4bf3daab90cd9},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11238. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Diffusion models have achieved remarkable success in image generation, yet tracing their outputs to individual training examples remains challenging. Existing attribution methods often compress factor-specific effects into scalar responses, making distinct internal changes indistinguishable. This is particularly limiting for diffusion models, where semantic factors emerge through evolving representation dynamics during denoising. We therefore reformulate diffusion data attribution as attributing factor-induced internal response trajectories. In this paper, we propose a novel Concept Attribution method through Dynamic Trajectories(CADT). We argue that attribution should therefore ask not only \emph{which} examples matter, but also \emph{how} their influence unfolds during generation. Specifically, we construct matched counterfactual pairs at identical noisy states to isolate factor-specific representation displacements, and model their directional and magnitude evolution across denoising as dynamic attribution signatures. For each training example and generated query, CADT extracts stage-wise feature vectors and integrates them along the denoising process to form a trajectory descriptor. Applying the same construction across the training set yields a bank of factor-specific trajectory descriptors. The covariance statistics of this bank are then used to construct . CADT uses this covariance-aware positive-semidefinite kernel to calibrate the query and training representations, and compares the calibrated query trajectory with each training trajectory to produce the final training-sample attribution scores. Experiments on multiple public datasets show consistent improvements over existing diffusion attribution baselines across hierarchical, compositional, and style attribution.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Graphics (cs.GR).
Pagina del lavoro: https://arxiv.org/abs/2610.11238 — 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 2026-10-09T05:55:07.286Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).