AI Paper Atlas

Research map / Continual learning / Online and replay-based

Lifelong Embodied Navigation Learning

Xudong Wang, Jiahua Dong, Baichen Liu, Qi Lyu, Lianqing Liu, Zhi Han

ICLR 2026 · Continual learning › Online and replay-based · approach: Plasticity-preserving continual reinforcement learning · 4 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 Inheritance Strategy (KIS) (ablation)Uni-Walker · LENL benchmarkSR (%): 60.3 -> 67.3Improved
add shared smoothing consolidation loss (SSC) (ablation)Uni-Walker · LENL benchmarkSR (%): 59.7 -> 67.3Improved
add Experts Co-Activation Strategy (ECAS) (ablation)Uni-Walker · LENL benchmarkSR (%): 58.1 -> 67.3Improved
add expert subspace orthogonality constraint (ESOC) (ablation)Uni-Walker · LENL benchmarkSR (%): 63.5 -> 67.3Improved
add Navigation Specific Chain of Thought (NSCoT) (ablation)Uni-Walker · LENL benchmarkSR (%): 51.1 -> 67.3Improved
use mixed matching instead of instruction-only matching in TAKA (ablation)Uni-Walker · LENL benchmarkSR (%): 35.0 -> 67.3Improved
use mixed matching instead of observation-only matching in TAKA (ablation)Uni-Walker · LENL benchmarkSR (%): 65.1 -> 67.3Improved
use mixed matching instead of observation-only matching in TAKA (ablation)Uni-Walker · LENL benchmarkSPL (%): 62.7 -> 62.3Hurt
add KIS (ablation)Uni-Walker · LENL benchmarkSR-F forgetting (%): 14.2 -> 4.3Improved
add ESOC (ablation)Uni-Walker · LENL benchmarkSR-F forgetting (%): 9.8 -> 4.3Improved
add TAKA to a LoRA continual learner (Seq-FT 12 vs O-LoRA+TAKA 58)LoRA continual learners · LENL benchmarkSR (%): 12 -> 58Improved
use Uni-Walker instead of previous best on unseen scenesSD-LoRA+TAKA · LENL unseen scenes S16-S18SR (%): 57.0 -> 62.0Improved

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