AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Improved Image Captioning via Policy Gradient Optimization of SPIDEr

Siqi Liu, Zhenhai Zhu, Ning Ye, Sergio Guadarrama, Kevin Murphy

ICCV 2017 · Vision-language and multimodal › Human and embodied modeling · approach: Neural image captioning · 465 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
optimize BCMR reward with policy gradient instead of MLEShow and Tell · COCO C-5 test (leaderboard)CIDEr-D: 0.947 -> 1.013Improved
optimize CIDEr alone with policy gradientShow and Tell · COCO C-5 test (leaderboard)CIDEr-D: 0.947 -> 0.995Improved
optimize BLEU-4 alone with policy gradientShow and Tell · COCO C-5 test (leaderboard)BLEU-4: 0.294 -> 0.346Improved
optimize BLEU-4 alone with policy gradientShow and Tell · COCO C-5 test (leaderboard)METEOR: 0.251 -> 0.249Hurt
use Monte Carlo rollout policy gradient instead of MIXER (SGD)Show and Tell · COCO C-5 test (leaderboard)CIDEr-D: 0.924 -> 1.013Improved
train MIXER with Adam instead of vanilla SGD (ablation)MIXER-BCMR · COCO C-5 test (leaderboard)CIDEr-D: 0.924 -> 0.991Improved
use rollout PG instead of MIXER with AdamShow and Tell · COCO C-5 test (leaderboard)METEOR: 0.258 -> 0.257No clear effect
optimize SPIDEr instead of BCMRShow and Tell · COCO C-5 test (leaderboard)CIDEr-D: 1.013 -> 1.000Hurt
optimize SPIDEr instead of BCMRShow and Tell · Human evaluation (492 COCO test images)fraction of not-bad captions: 4% marginImproved
train with policy gradient (BCMR or SPIDEr) instead of MLEShow and Tell · Human evaluation (492 COCO test images)fraction of not-bad captionsImproved
add CIDEr to SPICE reward (SPIDEr) instead of SPICE alone (ablation)Show and Tell · COCO hold-out validation captionscaption quality (qualitative)Improved
optimize BLEU-4 with MIXER instead of MLE (original MIXER split)MIXER · COCO 5000 validation imagesBLEU-4: 0.278 -> 0.292Improved

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