AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

SEVA: Leveraging sketches to evaluate alignment between human and machine visual abstraction

Kushin Mukherjee, Holly Huey, Xuanchen Lu, Yael Vinker, Rio Aguina-Kang, Ariel Shamir, Judith Fan

NeurIPS 2023 · Vision-language and multimodal › Human and embodied modeling · approach: Abstract visual reasoning benchmarks · 4 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase drawing time limit from 4 s to 32 sHuman sketches · SEVA human recognitiontop-1 accuracy: .031 -> .199Improved
increase drawing time limit from 4 s to 32 sHuman sketches · SEVA human recognitionresponse entropy: 1.958 -> 1.555Improved
increase drawing time limit from 4 s to 32 sHuman sketches · SEVA human recognitionsemantic neighbor preference: .628 -> .787Improved
more detailed sketches17 vision models · SEVAtop-1 accuracy: 9.57\times10^{-2}Improved
more detailed sketches17 vision models · SEVAresponse entropy: -.03Improved
more detailed sketches17 vision models · SEVAsemantic neighbor preference: 2.5 \times10^{-2}Improved
vision models compared with each other17 vision models · SEVAsemantic neighbor alignment: 5.05No clear effect
CLIPasso machine sketches vs human sketches at fewer strokesCLIPasso · SEVAJSD of human labels: -0.69Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy