AI Paper Atlas

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

Learning to Specialize with Knowledge Distillation for Visual Question Answering

Jonghwan Mun, Kimin Lee, Jinwoo Shin, Bohyung Han

NeurIPS 2018 · Vision-language and multimodal › Visual question answering · approach: Attention and prior-robust VQA · 31 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 knowledge distillation loss for non-assigned models in MCL (MCL-KD vs MCL) (ablation)MCL ensemble · CLEVR val, SAN, k=1top-1 accuracy: 42.19 -> 85.98Improved
replace uniform-distribution regularization (CMCL) with knowledge distillation (MCL-KD) (ablation)CMCL · CLEVR val, SAN, k=3top-1 accuracy: 86.18 -> 88.16Improved
MCL-KD instead of independent ensembleindependent ensemble · CLEVR val, SAN, k=1top-1 accuracy: 85.23 -> 85.98Improved
MCL-KD instead of independent ensembleindependent ensemble · CLEVR val, MLP, k=1top-1 accuracy: 60.10 -> 60.22Improved
MCL-KD instead of independent ensembleindependent ensemble · VQA v2.0 val, k=3top-1 accuracy: 65.27 -> 65.67Improved
CMCL instead of independent ensembleindependent ensemble · VQA v2.0 val, k=3top-1 accuracy: 65.27 -> 64.99No clear effect
increase number of specialized models per example k from 1 to 3 in MCL-KD (ablation)MCL-KD · CLEVR val, SANtop-1 accuracy: 85.98 -> 88.16Improved
MCL-KD instead of independent ensembleindependent ensemble · CIFAR-100, ResNet-20, k=1top-1 accuracy: 64.60 -> 65.27Improved
MCL-KD instead of independent ensembleindependent ensemble · CIFAR-100, VGGNet-17, k=1top-1 accuracy: 68.43 -> 68.75Improved
MCL and CMCL instead of independent ensembleindependent ensemble · CIFAR-100, ResNet-20, k=1top-1 accuracy: 64.60 -> 56.40 (MCL), 48.67 (CMCL)Hurt
add feature sharing to MCL-KD (ablation)MCL-KD · CIFAR-100, ResNet-20, k=1top-1 accuracy: 65.27 -> 66.72Improved
add feature sharing to CMCL (ablation)CMCL · CIFAR-100, ResNet-20, k=1top-1 accuracy: 48.67 -> 56.30Improved

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