AI Paper Atlas

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

LingoQA: Video Question Answering for Autonomous Driving

Ana-Maria Marcu, Long Chen, Jan Hünermann, Alice Karnsund, Benoit Hanotte, Prajwal Chidananda, Saurabh Nair, Vijay Badrinarayanan, Alex Kendall, Jamie Shotton, Elahe Arani, Oleg Sinavski

ECCV 2024 · Vision-language and multimodal › Visual question answering · approach: Vision-language model VQA evaluation

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 LingoQA fine-tuning to generic VQA pre-trained model (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 33.60 -> 60.80Improved
add generic VQA pre-training (GQA, SVIT) before fine-tuning (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 56.60 -> 60.80Improved
combine Action and Scenery fine-tuning datasets (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 55.40 -> 60.80Improved
use 5 video frames instead of a single frame (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 57.00 -> 60.80Improved
increase frames from 5 to 7 (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 60.80 -> 60.60No clear effect
use late fusion instead of early fusion (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 48.40 -> 60.80Improved
use late fusion instead of mid fusion (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 59.20 -> 60.80Improved
use Vicuna-1.5-7B instead of Mistral-7B-Instruct (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 58.00 -> 60.80Improved
use Vicuna-1.5-7B instead of OPT-7B (ablation)LingoQA Baseline · LingoQA evaluationLingo-Judge accuracy (%): 50.00 -> 60.80Improved
provide temporal context (5 frames instead of 1) to humansHuman evaluation · LingoQA evaluationLingo-Judge accuracy (%): 81.8 -> 96.6Improved
add visual input to fine-tuned LLM (LLaVA vs text-only Vicuna-7B)fine-tuned Vicuna-7B · LingoQA evaluationLingo-Judge accuracy (%): 38.8 -> 59.0Improved
use Lingo-Judge instead of GPT-4 as automatic metricevaluation metric · Correlation with human ratingsSpearman correlation: 0.941 -> 0.950Improved

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