AI Paper Atlas

Research map / Graph learning / Knowledge and temporal graphs

Lightweight Spatio-Temporal Modeling via Temporally Shifted Distillation for Real-Time Accident Anticipation

Patrik Patera, Yie-Tarng Chen, Wen-Hsien Fang

ICLR 2026 · Graph learning › Knowledge and temporal graphs · approach: Spatio-temporal graph traffic forecasting

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 contrastive video-text loss in pre-training (ablation)student with spatial and temporal distillation · DADmAP (%): 70.1 -> 75.3Improved
add temporally shifted distillation to spatial-only distillation (ablation)student · DADmAP (%): 71.2 -> 75.3Improved
add spatial distillation (ablation)student · DADmAP (%): 74.1 -> 75.3Improved
use a one-frame temporal shift instead of none (ablation)student · DADmAP (%): 71.2 -> 75.3Improved
increase temporal shift from 1 to 3 frames (ablation)student · DADmAP (%) and mTTA: mAP 75.3 -> 70.7; mTTA 4.04 -> 4.13Mixed
increase RWKV layers from 6 to 8 (ablation)student · DADmAP (%): 75.33 -> 71.6Hurt
use moderate 30% masking in memory strategy (ablation)student pre-training · DADmAP (%): 74.0 -> 75.3Improved
raise mask ratio to 50% or 75% (ablation)student pre-training · DADmAP (%): 75.3 -> 71.8 (50%), 71.0 (75%)Hurt
swap image teacher MobileCLIP for video teacher V-JEPA2 (ablation)distillation · DADmAP (%): 75.3 -> 66.0Hurt
add temporal shift when distilling from video teacher V-JEPA2 (ablation)distillation · DADmTTA (s): 4.12 -> 4.21Improved
add temporally shifted distillation to temporal RWKV (vs no pre-training) (ablation)student · DADmAP (%): 39.4 -> 75.3Improved
add temporal recurrence (T-RWKV) to spatial RWKV with TSD (ablation)student · DADmAP (%): 55.6 -> 75.3Improved

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