AutoScale is a closed-loop data engine using Graph-RAE for scene representation and Cluster-GA for importance-based retrieval to improve real-synthetic co-training for autonomous driving.
arXiv preprint arXiv:2506.08228 , year=
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
representative citing papers
World Engine generates realistic safety-critical driving variations from logs for reinforcement post-training, reducing benchmark failures more than data scaling and showing collision reductions plus on-road gains in a production system.
Self-play DAgger training in a batched pixel renderer produces end-to-end driving policies that reach competitive performance on HUGSIM and NAVSIM-v2 after real-world adaptation and improve with more self-play compute.
STELLAR trains up to 500M-parameter multi-modal models on 50M driving scenes and reports empirical scaling trends plus new state-of-the-art results on the Waymo Open Dataset.
VL-DPO uses a VLM as a zero-shot reasoner to generate preference pairs from pretrained model rollouts, then finetunes via DPO on the Waymo Open End-to-End Driving Dataset, yielding 11.94% higher rater feedback score and 10.01% lower average displacement error.
MOSAIC is a scaling-aware data selection framework that outperforms baselines in training end-to-end autonomous driving planners, achieving comparable or better EPDMS scores with up to 80% less data.
SimScale synthesizes unseen driving states from real logs via neural rendering and reactive environments, generates pseudo-expert trajectories, and shows that co-training on real plus simulated data improves planning robustness and generalization on real benchmarks, with gains scaling by simulation
DriveVLA-W0 adds world modeling to predict future images in VLA models, overcoming sparse action supervision and amplifying data scaling laws on NAVSIM benchmarks and a large in-house dataset.
DeepFleet develops and compares four foundation model architectures for multi-agent robot fleet coordination using warehouse data, finding robot-centric and graph-floor models most promising for prediction and scaling.
Autonomous driving policies with strong closed-loop performance frequently lack timely internal predictions of surrounding vehicles during near-collision events, and causal correction of prediction errors leads to improved ego planning.
Self-play RL regularized with 30 minutes of human data produces driving policies that coordinate with humans, training in 15 hours on one GPU with 2500x less data than imitation learning.
A differentiable motion forecasting model retrieves and refines interpretable trajectory anchors from a contrastively learned motion bank to improve transparency without sacrificing multi-modal accuracy.
Cross-benchmark analysis of 8 methods shows NAVSIM PDM Score correlates with Bench2Drive Driving Score at Spearman ρ=0.90, with Ego Progress as the strongest single predictor and a simpler 3-metric formula matching the full score.
citing papers explorer
-
Closed Loop Dynamic Driving Data Mixture for Real-Synthetic Co-Training
AutoScale is a closed-loop data engine using Graph-RAE for scene representation and Cluster-GA for importance-based retrieval to improve real-synthetic co-training for autonomous driving.
-
World Engine: Towards the Era of Post-Training for Autonomous Driving
World Engine generates realistic safety-critical driving variations from logs for reinforcement post-training, reducing benchmark failures more than data scaling and showing collision reductions plus on-road gains in a production system.
-
Scaling Self-Play for End-to-End Driving
Self-play DAgger training in a batched pixel renderer produces end-to-end driving policies that reach competitive performance on HUGSIM and NAVSIM-v2 after real-world adaptation and improve with more self-play compute.
-
STELLAR: Scaling 3D Perception Large Models for Autonomous Driving
STELLAR trains up to 500M-parameter multi-modal models on 50M driving scenes and reports empirical scaling trends plus new state-of-the-art results on the Waymo Open Dataset.
-
VL-DPO: Vision-Language-Guided Finetuning for Preference-Aligned Autonomous Driving
VL-DPO uses a VLM as a zero-shot reasoner to generate preference pairs from pretrained model rollouts, then finetunes via DPO on the Waymo Open End-to-End Driving Dataset, yielding 11.94% higher rater feedback score and 10.01% lower average displacement error.
-
Scaling-Aware Data Selection for End-to-End Autonomous Driving Systems
MOSAIC is a scaling-aware data selection framework that outperforms baselines in training end-to-end autonomous driving planners, achieving comparable or better EPDMS scores with up to 80% less data.
-
DriveVLA-W0: World Models Amplify Data Scaling Law in Autonomous Driving
DriveVLA-W0 adds world modeling to predict future images in VLA models, overcoming sparse action supervision and amplifying data scaling laws on NAVSIM benchmarks and a large in-house dataset.
-
DeepFleet: Multi-Agent Foundation Models for Mobile Robots
DeepFleet develops and compares four foundation model architectures for multi-agent robot fleet coordination using warehouse data, finding robot-centric and graph-floor models most promising for prediction and scaling.
-
What Probing Reveals about Autonomous Driving: Linking Internal Prediction Errors to Ego Planning
Autonomous driving policies with strong closed-loop performance frequently lack timely internal predictions of surrounding vehicles during near-collision events, and causal correction of prediction errors leads to improved ego planning.
-
Human-like autonomy emerges from self-play and a pinch of human data
Self-play RL regularized with 30 minutes of human data produces driving policies that coordinate with humans, training in 15 hours on one GPU with 2500x less data than imitation learning.
-
Recall to Predict: Grounding Motion Forecasting in Interpretable Motion Bank
A differentiable motion forecasting model retrieves and refines interpretable trajectory anchors from a contrastively learned motion bank to improve transparency without sacrificing multi-modal accuracy.