AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

DISCODE: Distribution-Aware Score Decoder for Robust Automatic Evaluation of Image Captioning

Nakamasa Inoue, Kanoko Goto, Masanari Oi, Martyna Gruszka, Mahiro Ukai, Takumi Hirose, Yusuke Sekikawa

AAAI 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Test-time adaptation of vision-language models · 0 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
replace token-probability score smoothing with ATT-loss Gaussian-prior decodingLLaVA-Next-72B evaluator · MCEval meanaccuracy: 82.5 -> 83.6Improved
add cross-entropy term to the ATT loss (ablation)DISCODE · MCEvalaccuracy: 81.8 -> 83.6Improved
add divergence term to the ATT loss (ablation)DISCODE · Flickr8k-ExpertKendall tau-c: 49.9 -> 56.1Improved
use adaptive alpha based on raw score instead of fixed (ablation)DISCODE · MCEvalaccuracy: 83.0 -> 83.6Improved
use continuous 0.0 to 1.0 scale instead of 0 to 9 (ablation)DISCODE · Flickr8k-ExpertKendall tau-c: 55.4 -> 56.1Improved
use numeric scale instead of letter scale A to E (ablation)DISCODE · Flickr8k-ExpertKendall tau-c: 54.4 -> 56.1Improved
use 1 to 5 scale instead of continuous scale (ablation)DISCODE · MCEvalaccuracy: 83.6 -> 83.6No clear effect
use weighted KL instead of standard KL divergence (ablation)DISCODE · MCEvalaccuracy: 83.0 -> 83.6Improved
use weighted KL instead of Jensen-Shannon divergence (ablation)DISCODE · MCEvalaccuracy: 81.6 -> 83.6Improved
apply DISCODE to InternVL-2.5-78B instead of FLEUR smoothingInternVL-2.5-78B · MCEvalaccuracy: 74.0 -> 78.1Improved
apply DISCODE to Qwen2-VL-7B instead of FLEUR smoothingQwen2-VL-7B-Instruct · MCEvalaccuracy: 75.2 -> 75.2No clear effect
apply DISCODE to MiniCPM-V-2.6 instead of FLEUR smoothingMiniCPM-V-2.6 · CompositeKendall tau-c: 66.2 -> 66.2No clear effect

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