AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

ExpertAD: Enhancing Autonomous Driving Systems with Mixture of Experts

Haowen Jiang, Xinyu Huang, You Lu, Dingji Wang, Yuheng Cao, Chaofeng Sha, Bihuan Chen, Keyu Chen, Xin Peng

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors · 1 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 Router (top-4 expert selection) to MoSE (ablation)Expert-UniAD · nuScenes open-loopAvg L2: 0.98 -> 0.86Improved
add Router (top-4 expert selection) to MoSE (ablation)Expert-UniAD · nuScenes open-loopAvg collision rate: 0.27 -> 0.21Improved
add sparse attention to MoSE (ablation)Expert-UniAD · nuScenes open-looplatency: 623 ms -> 476 msImproved
add sparse attention to Router-enabled MoSE (ablation)Expert-UniAD · nuScenes open-loopAvg L2: 0.86 -> 0.89Hurt
add sparse attention to Router-enabled MoSE (ablation)Expert-UniAD · nuScenes open-looplatency: 591 ms -> 445 msImproved
use Top-4 experts instead of all 8 (ablation)UniAD + MoSE · Bench2Drive closed-loopDriving Score: 46.24 -> 49.32Improved
use Top-4 experts instead of all 8 (ablation)UniAD + MoSE · Bench2Drive closed-loopRoute Completion: 67.54 -> 72.03Improved
increase PA tau from 32 to 128 (ablation)UniAD + PA · Bench2Drive closed-loopDriving Score: 40.35 -> 52.53Improved
increase PA tau from 128 to 256 (ablation)UniAD + PA · Bench2Drive closed-loopDriving Score: 52.53 -> 43.31Hurt
add addition function to PA with MLP (ablation)Expert-UniAD · nuScenes trackingAMOTA: 0.384 -> 0.404Improved
add MLP layer to PA with addition (ablation)Expert-UniAD · nuScenes trackingAMOTA: 0.390 -> 0.404Improved
add ExpertAD (PA + MoSE) to UniADUniAD · Bench2Drive closed-loopDriving Score: 44.62 -> 55.49Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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