AI Paper Atlas

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

A Deep Reinforcement Learning Based Multi-Step Coarse to Fine Question Answering (MSCQA) System

Yu Wang, Hongxia Jin

AAAI 2019 · Vision-language and multimodal › Visual question answering · approach: Attention and prior-robust VQA

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 action a3 (reject false-positive answers) and sub-context selection (ablation)MSCQA · WikiReadinganswer accuracy (%): 74.3 -> 75.8Improved
add action a3 (reject false-positive answers) and sub-context selection (ablation)MSCQA · WR-Longanswer accuracy (%): 42.8 -> 43.6Improved
add action a3 (reject false-positive answers) and sub-context selection (ablation)MSCQA · CNNanswer accuracy (%): 76.3 -> 77.6Improved
add action a3 (reject false-positive answers) and sub-context selection (ablation)MSCQA · SQuAD1.1 devanswer accuracy (%): 80.5 -> 83.6Improved
multi-step action selection instead of fixed single pipeline (Base Model 1)MSCQA · WikiReadingtraining speed (samples/sec): 112 -> 186Improved
multi-step action selection instead of fixed single pipeline (Base Model 1)MSCQA · WR-Longtraining speed (samples/sec): 66 -> 129Improved
multi-step action selection instead of fixed single pipeline (Base Model 1)MSCQA · CNNtraining speed (samples/sec): 83 -> 173Improved
multi-step action selection instead of fixed single pipeline (Base Model 1)MSCQA · SQuADtraining speed (samples/sec): 146 -> 213Improved
MSCQA vs QANet-style Base Model 2 on long documentsMSCQA · WR-Longanswer accuracy (%): 39.5 -> 43.6Improved
MSCQA vs coarse-to-fine Base Model 1 on short documentsMSCQA · SQuAD1.1 devanswer accuracy (%): 75.2 -> 83.6Improved
MSCQA vs strong baseline on CNNMSCQA · CNNanswer accuracy (%): 75.8 -> 77.6Improved
MSCQA vs Oracle sentence selectionMSCQA · SQuAD1.1 devanswer accuracy (%): 86.2 vs 83.6Hurt

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