AI Paper Atlas

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

TIME: Tensor-Factorized Mixture-of-Experts with Intrinsic Routing for Lifelong Multimodal Knowledge Editing

Dexuan Xu, Jieyi Wang, Shijie Li, Hanpin Wang, Yongzhi Cao, Yu Huang

ICML 2026 · Continual learning › Continual fine-tuning of LLMs · approach: Mixture-of-experts for multimodal continual tuning

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 score-threshold expert selection (vs using all experts) (ablation)TIME · E-VQA lifelong editing (BLIP2, gap=1000)average score: 90.17 -> 94.15Improved
add projection-score-based mixing weights (vs average) (ablation)TIME · E-VQA lifelong editing (BLIP2, gap=1000)average score: 86.67 -> 94.15Improved
add the structure-aware alignment loss Lalign (ablation)TIME · E-VQA lifelong editing (BLIP2, gap=1000)average score: 93.05 -> 94.15Improved
use CP decomposition instead of full-rank matrix experts (ablation)TIME · E-VQA lifelong editing (BLIP2, gap=1000)average score: 93.49 -> 94.15Improved
replace LiveEdit's dense experts and external routing with CP experts and intrinsic routingTIME · E-VQA (BLIP2)parameter size: 668 MB -> 523 MBImproved
replace LiveEdit's dense experts and external routing with CP experts and intrinsic routingTIME · E-VQA (LLaVA-v1.5)parameter size: 1178 MB -> 786 MBImproved
replace LiveEdit's external routing with intrinsic routingTIME · E-VQA (BLIP2, single editing)edit time (s): 239.25 -> 163.43Improved
replace LiveEdit's external routing with intrinsic routingTIME · E-VQA (BLIP2, gap=1000)edit time (s): 283.73 -> 178.92Improved
TIME vs LiveEdit on lifelong editingTIME · E-VQA (BLIP2, gap=1000)average score: 93.69 -> 94.15Improved
TIME vs LiveEdit on lifelong editingTIME · E-IC (BLIP2, gap=1000)average score: 82.22 -> 82.53Improved
TIME vs LiveEdit on single editingTIME · E-VQA (BLIP2, gap=1)average score: 96.94 -> 96.70No clear effect
increase module dimension H from 256 to 2048 (ablation)TIME · E-VQA lifelong editing (BLIP2, gap=1000)average scoreImproved

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