AI Paper Atlas

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

Learning to Align Question and Answer Utterances in Customer Service Conversation with Recurrent Pointer Networks

Shizhu He, Kang Liu, Weiting An

AAAI 2019 · Vision-language and multimodal › Visual question answering · approach: Open-domain multimodal QA · 10 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 separate customer and server encoders on top of shared conversation encoder (Model-II vs Model-I) (ablation)RPN (regression loss) · CS alignment overallF1 (%): 82.75 -> 84.86Improved
use separate customer and server encoders (Model-II vs Model-I) (ablation)RPN (classification loss) · CS alignment overallF1 (%): 82.82 -> 83.49Improved
use regression loss instead of classification loss (ablation)RPN Model-II · CS alignment One-to-ManyF1 (%): 76.38 -> 80.39Improved
use regression loss instead of classification loss (ablation)RPN Model-I · CS alignment One-to-ManyF1 (%): 69.46 -> 78.37Improved
use regression loss instead of classification loss (ablation)RPN Model-I · CS alignment overallF1 (%): 82.82 -> 82.75No clear effect
use regression loss instead of classification loss (ablation)RPN Model-II · CS alignment overallF1 (%): 83.49 -> 84.86Improved
replace content matching with recurrent pointer network alignmentQ-A alignment · CS alignment overallF1 (%): 49.28 -> 84.86Improved
replace pairwise two-stage method with RPNSHCNN+CISIR · CS alignment overallF1 (%): 79.5 -> 84.86Improved
use RPN instead of SHCNN+CISIRSHCNN+CISIR · CS alignment None alignmentsF1 (%): 83.32 -> 94.07Improved
add threshold strategy to content matchingMatch-G (CNN) · CS alignment overallF1 (%): 49.28 -> 46.49Hurt
remove conversation-encoder outputs from inputs of customer/server encoders (ablation)RPN Model-II · CS alignment overallF1 (%)Hurt

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