AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

UniTraj: A Unified Framework for Scalable Vehicle Trajectory Prediction

Lan Feng, Mohammadhossein Bahari, Kaouther Messaoud, Eloi Zablocki, Matthieu Cord, Alexandre Alahi

ECCV 2024 · Probabilistic and scientific modeling › Molecules and proteins · approach: Multimodal trajectory prediction · 62 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
train on all combined datasets instead of nuScenes onlyMTR · nuScenes valbrier-minFDE: 2.86 -> 2.27Improved
train on all combined datasets instead of WOMD onlyMTR · WOMD valbrier-minFDE: 2.13 -> 2.13No clear effect
evaluate models on a different dataset than they were trained onMTR, Wayformer, AutoBot · cross-datasetbrier-minFDE: 2.08 -> 4.50Hurt
evaluate AutoBot on a different city than trained onAutoBot · Cross-city nuPlanbrier-minFDE: 2.4 -> 3.5Hurt
control training size to 30k samples per dataset (ablation)AutoBot · cross-dataset averagebrier-minFDE: WOMD 3.89 vs Argoverse 2 4.21 vs nuScenes 4.91No clear effect
multi-dataset training at equal 30k sample size (ablation)AutoBot · 30k per dataset averagebrier-minFDE: 3.89 -> 3.06Improved
increase combined training data from 20% to 100% (ablation)AutoBot · average over three datasetsbrier-minFDEImproved
combined data vs larger-capacity modelsAutoBot (1.5M) vs MTR (60.1M) · nuScenes valbrier-minFDE: AutoBot 3.36 -> 3.07; MTR 2.86 -> 2.27Mixed
train MTR on combined data, medium Kalman difficulty samples (ablation)MTR · nuScenes valbrier-minFDE: 4.52 -> 2.97Improved
train MTR on combined data, hard Kalman difficulty samples (ablation)MTR · nuScenes valbrier-minFDE: 4.25 -> 4.20No clear effect
fine-tune WOMD-trained model on Argoverse 2 (continual learning)AutoBot · WOMD and Argoverse 2 valbrier-minFDE: +0.92 improvement, -0.52 forgettingMixed
add synthetic PG data to nuScenes trainingAutoBot · nuScenes valbrier-minFDE: 3.55 -> 3.56No clear effect

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