AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Forecast-MAE: Self-supervised Pre-training for Motion Forecasting with Masked Autoencoders

Jie Cheng, Xiaodong Mei, Ming Liu

ICCV 2023 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Masked autoencoder pretraining · 125 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 agents' future trajectories as input to lane-masked pre-training (ablation)Forecast-MAE · Argoverse 2 valminADE6: 0.865 -> 0.828Improved
mask history only with future given unmasked (ablation)Forecast-MAE · Argoverse 2 valminADE6: 0.8314 -> 0.864Hurt
add future masking (complementary with history) (ablation)Forecast-MAE · Argoverse 2 valminADE6: 0.866 -> 0.814Improved
increase encoder depth from 2 to 4 (ablation)Forecast-MAE · Argoverse 2 valminADE6: 0.854 -> 0.814Improved
increase encoder depth from 4 to 5 (ablation)Forecast-MAE · Argoverse 2 valminADE6: 0.814 -> 0.815No clear effect
add self-supervised pre-training before fine-tuningForecast-MAE · Argoverse 2 valminADE6: 0.811 -> 0.801Improved
train from scratch for more epochs (60 to 80) (ablation)Forecast-MAE scratch · Argoverse 2 valminADE6: 0.811 -> 0.814No clear effect
add pre-training under distribution shift between city groups (ablation)Forecast-MAE · Argoverse 2 (train Miami/Pittsburgh/Austin, test other cities)minADE: 0.9098 -> 0.8968Improved
add MAE pre-training vs SSL-Lanes pretext tasksForecast-MAE vs SSL-Lanes · Argoverse 2 valminADE6: 0.850 -> 0.801Improved
add ensemble of 6 variants (ablation)Forecast-MAE · Argoverse 2 testminFDE1: 4.355 -> 4.145Improved
add pre-training on Waymo Open Motion DatasetForecast-MAE · WOMD testminADE: 0.689 -> 0.632Improved
add pre-training on Argoverse 1Forecast-MAE · Argoverse 1 valminADE: 0.736 -> 0.710Improved

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