AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Progressive Pretext Task Learning for Human Trajectory Prediction

Xiaotong Lin, Tianming Liang, Jianhuang Lai, Jian-Fang Hu

ECCV 2024 · Probabilistic and scientific modeling › Molecules and proteins · approach: Multimodal trajectory prediction · 38 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 Task-I stepwise next-position pretraining (ablation)PPT · SDDdestination minFDE: 11.58 -> 10.70Improved
add Task-I and Task-II pretext tasks to the training pipeline (ablation)PPT · SDDminADE/minFDEImproved
add cross-task knowledge distillation (ablation)PPT · SDDminADE/minFDE (20 seeds)Improved
set diversity loss weight too small or too large instead of 100 (ablation)PPT Task-II · SDDminADE/minFDEHurt
replace autoregressive decoding with destination-driven two-step prediction and learnable promptsTransformer predictor · Inference speedinference time: 5.28ms vs 4.06ms (TUTR)Improved
PPT versus best prior method on SDDtrajectory prediction · SDDminADE20: 7.61 -> 7.03Improved
PPT versus best prior method on GCStrajectory prediction · GCSminADE20: 7.42 -> 6.20Improved

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