AI Paper Atlas

Research map / Vision-language and multimodal / Visual question answering

Recursive Visual Programming

Jiaxin Ge, Sanjay Subramanian, Baifeng Shi, Roei Herzig, Trevor Darrell

ECCV 2024 · Vision-language and multimodal › Visual question answering · approach: Universal VQA models · 9 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
add recursive_query (recursive code generation) to ViperGPTViperGPT · VSR zero-shot splitaccuracy: 61.59 -> 66.09Improved
add recursive_query to ViperGPTViperGPT · GQA test-devaccuracy: 44.63 -> 45.62Improved
specify return type explicitly in recursive_query (dynamic-explicit vs non-recursive) (ablation)RVP · GQA (recursive-pattern instances)accuracy: 55.99 -> 70.23Improved
let recursive_query return varying types without specifying them in the query (implicit) (ablation)RVP · COVR val_1000 (recursive-pattern instances)accuracy: 60.71 -> 53.57Hurt
fixed-type (string) recursive_query vs non-recursive (ablation)RVP · GQA vs COVR val_1000accuracy: 55.99 -> 65.05 (GQA); 60.71 -> 60.12 (COVR)Mixed
use retrieval-based selection of recursive in-context examples (ablation)RVP · GQA val_2000accuracy: 54.75 -> 53.79No clear effect
add or remove one in-context example (ablation)RVP · COVR val_1000accuracy: 51.05 -> 51.35 (add 1)No clear effect
replace gpt-3.5-turbo with CodeLlama-13B as code generator (ablation)RVP · GQA val_2000accuracy: 54.75 -> 49.27Hurt
apply NLP prompting methods (CoT, PoT, sub-functions) instead of recursion (ablation)visual programming · GQA val_2000accuracy: 54.75 (RVP) vs 48.42-52.03Hurt
recursive decomposition on more complex GQA questionsRVP · GQA 2-step relatesaccuracy: 32.37 -> 40.29 (+7.92)Improved
recursive decomposition on single-property GQA questionsRVP · GQA 1-step propertiesaccuracy: 64.62 -> 62.45 (-2.17)Hurt
recursive code on Compare questions in GQARVP · GQA test-dev (Compare)accuracy: 66.67 -> 53.33Hurt

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