AI Paper Atlas

Research map / Continual learning / Continual fine-tuning of LLMs

ECA: Efficient Continual Alignment for Open-Ended Image-to-Text Generation

Jiangtao Kong, Peijun Zhao, Chun-Fu Chen, Youngwook Do, Shaohan Hu, Tianyi Zhou, Huajie Shao

ICML 2026 · Continual learning › Continual fine-tuning of LLMs · approach: Mixture-of-experts for multimodal continual tuning · 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
add Mixture of Query (shared orthogonal task queries) to a parallel adapter (ablation)Vanilla (PA) on BLIP-2 · ToS-COCO CaptionCIDEr (Avg): 123.00 -> 123.67Improved
use per-task query tokens without cross-task sharing instead of MoQ (ablation)Vanilla (PA) on BLIP-2 · ToS-COCO CaptionCIDEr (Avg): 123.00 -> 122.74No clear effect
add Dictionary Replay on top of MoQ (ablation)PA+MoQ on BLIP-2 · ToS-COCO CaptionCIDEr (Avg): 123.67 -> 124.57Improved
replay a randomly initialized dictionary instead of a learned one (ablation)PA+MoQ+DR on BLIP-2 · ToS-COCO CaptionCIDEr (Avg): 124.57 -> 123.24Hurt
add Fisher Dynamic Expansion on top of MoQ (ablation)PA+MoQ on BLIP-2 · ToS-COCO CaptionCIDEr (Avg): 123.67 -> 124.95Improved
add Dictionary Replay to MoQ+FeDEx (full ECA) (ablation)PA+MoQ+FeDEx on BLIP-2 · ToS-COCO CaptionCIDEr (Avg): 124.95 -> 125.56Improved
use orthogonality loss only in MoQ (ablation)PA+MoQ on BLIP-2 · ToS-COCO CaptionCIDEr (BWT): -4.50 -> -3.99Improved
use key-alignment loss only in MoQ (no orthogonality) (ablation)PA+MoQ on BLIP-2 · ToS-COCO CaptionBLEU-4 (BWT): -1.49 -> -1.80Hurt
set DR weight lambda to 1 instead of 0.1 (ablation)PA+MoQ+DR · ToS-COCO CaptionCIDEr (Avg): 124.57 -> 122.84Hurt
use 2.5x dv dictionary atoms instead of 5x (ablation)PA+MoQ+DR · ToS-COCO CaptionCIDEr (Avg): 124.57 -> 122.42Hurt
set FeDEx expansion threshold S below 0.5 (ablation)ECA · ToS-COCO CaptionAvg metricsHurt
add EWC regularization to a single parallel adapterVanilla (PA) on BLIP-2 · ToS-TextCapsCIDEr (Avg): 89.39 -> 86.98Hurt

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