Paper Detail

Evaluating Cultural Knowledge Editing in Vision-Language Models

Malvina Nikandrou, Georgios Pantazopoulos, Ioannis Konstas, Alessandro Suglia

arxiv Score 11.3

Published 2026-10-29 · First seen 2026-10-30

General AI

Abstract

Current Vision-Language Models (VLMs) frequently fail to recognize culture-specific concepts, especially from underrepresented regions. Effectively injecting this long-tail knowledge into existing models remains a challenge. In this work, we investigate knowledge editing as a targeted alternative by introducing CulturalBind, a benchmark comprising 3.2K images of 1K concepts from 6 domains, designed to evaluate culture-aware cross-modal knowledge as well as concept-level knowledge editing in VLMs. Pre-edit evaluation reveals that textual and visual knowledge is decoupled: models often demonstrate text-based knowledge of cultural concepts but fail to identify them from images. We then evaluate editing methods in single and sequential scenarios. Although current methods successfully correct individual visual recognition errors, they exhibit low cross-modal portability, failing to integrate edits into the model's broader parametric knowledge. Furthermore, performance degrades rapidly under sequential editing, highlighting the need for comprehensive adaptation approaches that better consolidate multimodal knowledge over time.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
now
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{nikandrou2026evaluating,
  title = {Evaluating Cultural Knowledge Editing in Vision-Language Models},
  author = {Malvina Nikandrou and Georgios Pantazopoulos and Ioannis Konstas and Alessandro Suglia},
  year = {2026},
  abstract = {Current Vision-Language Models (VLMs) frequently fail to recognize culture-specific concepts, especially from underrepresented regions. Effectively injecting this long-tail knowledge into existing models remains a challenge. In this work, we investigate knowledge editing as a targeted alternative by introducing CulturalBind, a benchmark comprising 3.2K images of 1K concepts from 6 domains, designed to evaluate culture-aware cross-modal knowledge as well as concept-level knowledge editing in VLMs},
  url = {https://arxiv.org/abs/2610.39602},
  keywords = {cs.CV, cs.CL},
  eprint = {2610.39602},
  archiveprefix = {arXiv},
}

Metadata

{}