AI Paper Atlas

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

NexusFlow: Unifying Disparate Tasks under Partial Supervision via Invertible Flow Networks

Fangzhou Lin, Yuping Wang, Yuliang Guo, Zixun Huang, Xinyu Huang, Haichong Zhang, Kazunori Yamada, Zhengzhong Tu, Liu Ren, Ziming Zhang

CVPR 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 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 invertible coupling layers (ablation)NexusFlow · nuScenes PS-MTLAMOTA: 0.214 -> 0.329Improved
w/o inv vs baseline on tracking (ablation)NexusFlow · nuScenes PS-MTLAMOTA: 0.289 -> 0.214Hurt
increase coupling depth from 4 to 6 layers (ablation)NexusFlow · nuScenes PS-MTLAMOTA: 0.292 -> 0.329Improved
increase coupling depth from 6 to 8 layers (ablation)NexusFlow · nuScenes PS-MTLAMOTA: 0.329 -> 0.247Hurt
add NexusFlow alignment to UniAD (ablation)UniAD · nuScenes PS-MTLAMOTA: 0.289 -> 0.329Improved
add NexusFlow alignment to UniAD (ablation)UniAD · nuScenes PS-MTLLanes IoU (%): 27.1 -> 37.1Improved
add NexusFlow alignment to UniAD (ablation)UniAD · nuScenes PS-MTLDrivable IoU (%): 62.7 -> 64.5Mixed
add prior PS-MTL methods (MTPSL, JTR) designed for homogeneous tasks (ablation)UniAD · nuScenes PS-MTLAMOTA: 0.289 -> 0.255 (MTPSL), 0.197 (JTR)Hurt
two-stage fine-tuning instead of one-phase joint training (ablation)NexusFlow · nuScenes PS-MTLAMOTA: 0.318 -> 0.329Improved
center-based instead of pairwise alignment loss (ablation)NexusFlow · NYUv2Segmentation IoU: 31.35 -> 31.70No clear effect
add NexusFlow to MTL on NYUv2 (ablation)MTL baseline · NYUv2Segmentation IoU: 26.75 -> 31.70Improved
add NexusFlow vs MTPSL on memoryNexusFlow · NYUv2GPU memory: 8900 MiB -> 4554 MiBImproved

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