Adaptive-WAM: Quality-Guided Early-Exit Planning from Intermediate Video-Diffusion Features
Overview
This paper presents Adaptive-WAM, a planning framework built on a Wan2.2-5B video diffusion backbone that uses intermediate-layer features to decode trajectories and terminates inference early when quality thresholds are met. The method avoids iterative future-video synthesis, achieving 90.8 PDMS on NAVSIM with adaptive routing, 89.9 EPDMS on NAVSIM v2, and transferring to nuScenes with 0.88 m average L2 error and a 0.08% collision rate without fine-tuning. On an A100 GPU, the adaptive planner averages 170 ms end-to-end latency, improving over fixed-depth baselines while maintaining strong performance. Source code will be released.
Generated from attributed reports · Updated 3 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Reported quantity · PDMS score: 90.8 other · Basis not reported · Differing source assertions
“ng to trajectory quality. On NAVSIM, the adaptive single-trajectory planner achieves 90.8 PDMS; a separate fixed-exit variant reaches 92.6 PDMS with 64 proposals. It further obtains 89.9 EPDMS on NAVSIM v2, yielding the best reported results among the compared front-view video world-model planners. Without target-domain fine-tuning, Adaptive-WAM transfers to nuScenes with 0.88 m average L2 error and a 0.08\% collisio”
Exact source · revision 1Source owner not reported
Reported quantity · PDMS score: 92.6 other · Basis not reported · Differing source assertions
“ng to trajectory quality. On NAVSIM, the adaptive single-trajectory planner achieves 90.8 PDMS; a separate fixed-exit variant reaches 92.6 PDMS with 64 proposals. It further obtains 89.9 EPDMS on NAVSIM v2, yielding the best reported results among the compared front-view video world-model planners. Without target-domain fine-tuning, Adaptive-WAM transfers to nuScenes with 0.88 m average L2 error and a 0.08\% collisio”
Exact source · revision 1Source owner not reported
Reported quantity · EPDMS score: 89.9 other · Basis not reported
“ng to trajectory quality. On NAVSIM, the adaptive single-trajectory planner achieves 90.8 PDMS; a separate fixed-exit variant reaches 92.6 PDMS with 64 proposals. It further obtains 89.9 EPDMS on NAVSIM v2, yielding the best reported results among the compared front-view video world-model planners. Without target-domain fine-tuning, Adaptive-WAM transfers to nuScenes with 0.88 m average L2 error and a 0.08\% collisio”
Exact source · revision 1Source owner not reported
Reported quantity · average L2 error: 0.88 other · Basis not reported
“ng to trajectory quality. On NAVSIM, the adaptive single-trajectory planner achieves 90.8 PDMS; a separate fixed-exit variant reaches 92.6 PDMS with 64 proposals. It further obtains 89.9 EPDMS on NAVSIM v2, yielding the best reported results among the compared front-view video world-model planners. Without target-domain fine-tuning, Adaptive-WAM transfers to nuScenes with 0.88 m average L2 error and a 0.08\% collisio”
Exact source · revision 1Source owner not reported
Reported quantity · planning latency: 170 other · Basis not reported · Differing source assertions
“n rate. On an A100, adaptive routing improves PDMS from 90.62 to 90.79 while averaging 170 ms end-to-end planning latency, approximately 10\% below the 190 ms fixed block-15 planner and 47\% below the 320 ms fixed full-depth planner. Code will be released.”
Exact source · revision 1Source owner not reported
Reported quantity · planning latency: 190 other · Basis not reported · Differing source assertions
“n rate. On an A100, adaptive routing improves PDMS from 90.62 to 90.79 while averaging 170 ms end-to-end planning latency, approximately 10\% below the 190 ms fixed block-15 planner and 47\% below the 320 ms fixed full-depth planner. Code will be released.”
Exact source · revision 1Source owner not reported
Reported quantity · planning latency: 320 other · Basis not reported · Differing source assertions
“n rate. On an A100, adaptive routing improves PDMS from 90.62 to 90.79 while averaging 170 ms end-to-end planning latency, approximately 10\% below the 190 ms fixed block-15 planner and 47\% below the 320 ms fixed full-depth planner. Code will be released.”
Exact source · revision 1Source owner not reported
Artifact availability · code: reported pending
“n rate. On an A100, adaptive routing improves PDMS from 90.62 to 90.79 while averaging 170 ms end-to-end planning latency, approximately 10\% below the 190 ms fixed block-15 planner and 47\% below the 320 ms fixed full-depth planner. Code will be released.”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalAdaptive-WAM: Quality-Guided Early-Exit Planning from Intermediate Video-Diffusion Features
This paper presents Adaptive-WAM, a planning framework built on a Wan2.2-5B video diffusion backbone that uses intermediate-layer features to decode trajectories and terminates inference early when quality thresholds are met. The method avoids iterative future-video synthesis, achieving 90.8 PDMS on NAVSIM with adaptive routing, 89.9 EPDMS on NAVSIM v2, and transferring to nuScenes with 0.88 m average L2 error and a 0.08% collision rate without fine-tuning. On an A100 GPU, the adaptive planner averages 170 ms end-to-end latency, improving over fixed-depth baselines while maintaining strong performance. Source code will be released.
Event coverage history
Current attention 7·Peak within the comparable range 7（2026-11-02 18:00 UTC）·Change within the comparable range over 24 hours –
The trend compares participants observed continuously; its scope may be narrower than current attention totals. Hover or click to explore hourly observations, or use the left and right arrow keys.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).