AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

Learning Group Interactions and Semantic Intentions for Multi-Object Trajectory Prediction

Mengshi Qi, Yuxin Yang, Huadóng Ma

TPAMI 2026 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Interactive multi-agent trajectory prediction · 4 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
use diffusion framework instead of MLP encoder-decoder (ablation)trajectory predictor · NBA SportVUminADE/minFDE at 4.0s: 5.28/8.71 -> 0.84/1.17Improved
use diffusion framework instead of learnable-query attention (ablation)trajectory predictor · NBA SportVUminADE/minFDE at 4.0s: 1.24/1.63 -> 0.84/1.17Improved
add Banzhaf Interaction module (team-level) (ablation)tactic prediction · NBA SportVUTop-1 accuracy: 58.44 -> 60.09Improved
add Banzhaf Interaction module (scene-level) (ablation)tactic prediction · NBA SportVUTop-1 accuracy: 58.44 -> 60.23Improved
use scene-level instead of team-level Banzhaf learner (ablation)tactic prediction · NBA SportVUTop-1 accuracy: 60.09 -> 60.23Improved
use group tactic conditioning and Banzhaf intention vs LEDtrajectory predictor · NBA SportVUminADE at 4.0s: 0.89 -> 0.84Improved
use method vs LED on soccertrajectory predictor · TeamTrack-SoccerminADE/minFDE at 4.8s: 1.75/3.01 -> 1.48/2.72Improved
use method vs LED on basketballtrajectory predictor · TeamTrack-BasketballminADE at 4.8s: 2.01 -> 1.72Improved
add one-hot tactic embedding to baselineTransformer tactic predictor · NBA SportVUTop-1 accuracy: 34.56 -> 53.31Improved
add team pooling to tactic-embedding baselineTransformer tactic predictor · NBA SportVUTop-1 accuracy: 53.31 -> 55.22Improved
LoRA finetune Llama3-8B on the benchmarkLlama3-8B · NBA SportVUTop-1 accuracy: 22.39 -> 38.17Improved
use proposed method instead of finetuned Llama3-8Btactic predictor · NBA SportVUTop-1 accuracy: 38.17 -> 60.23Improved

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