Research map / Multi-agent systems and games / Human-AI and goal-conditioned
INSTINCT: Instance-Level Interaction Architecture for Query-Based Collaborative Perception
Yunjiang Xu, Lingzhi Li, Jin Wang, Yupeng Ouyang, Benyuan Yang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add quality-aware filtering to late-fusion baseline (ablation) | INSTINCT · DAIR-V2X | AP@0.7: 0.5978 -> 0.6041 | Improved |
| add quality-aware filtering to late-fusion baseline (ablation) | INSTINCT · DAIR-V2X | AP@0.5: 0.7302 -> 0.6962 | Hurt |
| add dual-branch detection routing (ablation) | INSTINCT · DAIR-V2X | AP@0.7: 0.6041 -> 0.6321 | Improved |
| add cross-domain adaption (ablation) | INSTINCT · DAIR-V2X | AP@0.7: 0.6321 -> 0.7101 | Improved |
| add Gaussian distance attention (ablation) | INSTINCT · DAIR-V2X | AP@0.7: 0.7101 -> 0.7394 | Improved |
| add Co-GT Sampling (ablation) | INSTINCT · DAIR-V2X | AP@0.7: 0.7394 -> 0.7529 | Improved |
| add MAL loss quality penalty (ablation) | INSTINCT · DAIR-V2X | AP@0.7: 0.7415 -> 0.7529 | Improved |
| add separate IoU quality prediction branch on top of MAL (ablation) | INSTINCT · DAIR-V2X | AP@0.7: 0.7529 -> 0.7401 | Hurt |
| add separate IoU quality prediction branch on top of MAL (ablation) | INSTINCT · V2XSet | AP@0.7: 0.8731 -> 0.8822 | Improved |
| replace ConQueR with weaker Voxel-DETR detector (ablation) | INSTINCT · DAIR-V2X | AP@0.7: 0.7529 -> 0.7303 | Hurt |
| instance-level INSTINCT versus Where2comm | INSTINCT · V2V4Real | AP@0.7: 0.3801 -> 0.6196 | Improved |
| instance-level INSTINCT versus Where2comm | INSTINCT · DAIR-V2X | AP@0.7: 0.6649 -> 0.7529 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Where2comm — Where2comm: Communication-Efficient Collaborative Perception via Spatial Confidence Maps (NeurIPS 2022) · better on 6 of 7 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- IFTR: An Instance-Level Fusion Transformer for Visual Collaborative Perception (ECCV 2024 · similarity 0.92)
- An Extensible Framework for Open Heterogeneous Collaborative Perception (ICLR 2024 · similarity 0.90)
- CoSDH: Communication-Efficient Collaborative Perception via Supply-Demand Awareness and Intermediate-Late Hybridization (CVPR 2025 · similarity 0.89)
- RCP-Bench: Benchmarking Robustness for Collaborative Perception Under Diverse Corruptions (CVPR 2025 · similarity 0.88)
- STAMP: Scalable Task- And Model-agnostic Collaborative Perception (ICLR 2025 · similarity 0.88)
- CoPEFT: Fast Adaptation Framework for Multi-Agent Collaborative Perception with Parameter-Efficient Fine-Tuning (AAAI 2025 · similarity 0.87)
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.
