Research map / Probabilistic and scientific modeling / Molecules and proteins
Deep Multi-Task Learning for Joint Localization, Perception, and Prediction
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Jointly add learned localization to the perception-prediction system | ILVM-based autonomy stack · LP3, noisy P2 pose and noisy planning pose | ℓ2 human (m @ 5s): 4.70 -> 4.64 | Helped |
| Jointly add learned localization to the perception-prediction system | ILVM-based autonomy stack · LP3, noisy P2 pose and noisy planning pose | collision rate (%): 3.511 -> 2.962 | Helped |
| Use the larger Pixor localization embedding instead of Tiny Pixor (ablation) | Joint LP2 · LP3, noisy P2 pose and noisy planning pose | recall @ 1 (%): 46.6 -> 52.5 | Helped |
| Use the larger Pixor localization embedding instead of Tiny Pixor (ablation) | Joint LP2 · LP3, noisy P2 pose and noisy planning pose | recall @ 2 (%): 93.5 -> 96.9 | Helped |
| Reduce online localization embedding size (ablation) | Localization network · Localization inference | inference time: four-fold | Helped |
| Reduce localization network size (ablation) | Localization network · Localization matching | matching accuracy | Hurt |
| Reuse perception feature maps in the localization embedding (ablation) | Joint LP2 · Localization inference | inference time (ms): 2.79 -> 1.95 | Helped |
| Add translational localization jitter up to 25 cm | ILVM perception-prediction · P2 under localization perturbation | mAP@0.7 | No clear effect |
| Add rotational localization jitter up to 0.5 degrees | ILVM perception-prediction · P2 under localization perturbation | mAP@0.7 | No clear effect |
| Increase localization error, especially rotational error | P2 and motion planning stack · P2 and planning under localization perturbation | system performance: ~ 7% mAP, .30 mean SFDE | Hurt |
Closest papers by meaning
- M3Net: Multimodal Multi-task Learning for 3D Detection, Segmentation, and Occupancy Prediction in Autonomous Driving (AAAI 2025 · similarity 0.86)
- DualAD: Disentangling the Dynamic and Static World for End-to-End Driving (CVPR 2024 · similarity 0.86)
- Prioritizing Perception-Guided Self-Supervision: A New Paradigm for Causal Modeling in End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.85)
- SpaceDrive: Infusing Spatial Awareness into VLM-based Autonomous Driving (CVPR 2026 · similarity 0.85)
- SGDrive: Scene-to-Goal Hierarchical World Cognition for Autonomous Driving (CVPR 2026 · similarity 0.85)
- Think Before You Drive: World Model-Inspired Multimodal Grounding (CVPR 2026 · similarity 0.85)
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.