AI Paper Atlas

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

VisoGender: A dataset for benchmarking gender bias in image-text pronoun resolution

Siobhan Mackenzie Hall, Fernanda Gonçalves Abrantes, Hanwen Zhu, Grace Sodunke, Aleksandar Shtedritski, Hannah Rose Kirk

NeurIPS 2023 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded scene text and structure · 3 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
Use captioning models instead of vision-language encoders for pronoun resolutionVLMs on VisoGender · VisoGender resolutionresolution accuracy: CLIP 0.75; BLIP-2 0.84; GIT 0.84Improved
Resolve pronouns in scenes with two people of different perceived genders rather than single-person scenesVLMs on VisoGender · VisoGender resolutionresolution accuracy: CLIP: 0.92 -> 0.36; BLIP-2: 0.92 -> 0.60Hurt
Use gender-neutral occupational queries for image retrievalVision-language encoders on VisoGender · VisoGender retrieval, two-person imagesBias@5: 0.11; 0.10; 0.17; 0.06; 0.11; 0.01Mixed
Reverse occupation and participant order in the prompt (ablation)CLIP on two-person pronoun resolution · VisoGender resolutionresolution accuracyHurt
Reverse occupation and participant order in the retrieval template (ablation)SLIP on VisoGender retrieval · VisoGender retrievalBias@5: 0.06 -> -0.18Mixed
Include a neutral pronoun option (ablation)CLIP on VisoGender · VisoGender neutral pronoun resolutionneutral pronoun resolution rate: 0.17 overall; 0.31 for two people with different perceived gendersNo clear effect
Include a neutral pronoun option (ablation)BLIP-2 on VisoGender · VisoGender neutral pronoun resolutionneutral pronoun resolution rate: 0.01 overall; 0.00 for two people with different perceived gendersNo clear effect

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