Research map / Continual learning / Online and replay-based
Deconfounded Lifelong Learning for Autonomous Driving via Dynamic Knowledge Spaces
Jiayuan Du, Yuebing Song, Yiming Zhao, Xianghui Pan, Jiawei Lian, Yuchu Lu, Liuyi Wang, Chengju Liu, Qijun Chen
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 evolutionary trajectory decoder (ablation) | DeLL · Bench2Drive lifelong learning | Avg driving score (%): 72.94 -> 74.69 | Improved |
| add evolutionary trajectory decoder (ablation) | DeLL · Bench2Drive lifelong learning | Backward transfer BT (%): 72.21 -> 79.63 | Improved |
| add evolutionary trajectory decoder (ablation) | DeLL · Bench2Drive lifelong learning | Forgetting ratio FR (%): 33.12 -> 33.97 | Hurt |
| add TFEM (trajectory front-door enhancement) (ablation) | DeLL · Bench2Drive lifelong learning | Avg driving score (%): 73.00 -> 74.69 | Improved |
| add TFEM (ablation) | DeLL · Bench2Drive lifelong learning | Process forgetting ratio PFR (%): 32.86 -> 29.80 | Improved |
| add TFEM (ablation) | DeLL · Bench2Drive lifelong learning | Forward transfer FT (%): 40.04 -> 42.88 | Improved |
| add FFEM (fused feature front-door enhancement) (ablation) | DeLL · Bench2Drive lifelong learning | Forgetting ratio FR (%): 38.33 -> 33.97 | Improved |
| add FFEM (ablation) | DeLL · Bench2Drive lifelong learning | Forward transfer FT (%): 36.59 -> 42.88 | Improved |
| add FFEM (ablation) | DeLL · Bench2Drive lifelong learning | Avg driving score (%): 73.10 -> 74.69 | Improved |
| apply DeLL (dual knowledge spaces, causal enhancement, evolutionary decoder) to TF++ | TF++ · Bench2Drive lifelong learning | Forgetting ratio FR (%): 44.50 -> 33.97 | Improved |
| apply DeLL to TF++ | TF++ · Bench2Drive lifelong learning | Backward transfer BT (%): 52.83 -> 79.63 | Improved |
| apply DeLL to TF++ | TF++ · Bench2Drive lifelong learning | Avg driving score (%): 70.55 -> 74.69 | Improved |
Closest papers by meaning
- Hybrid-Driving: An Autonomous Driving Decision Framework Integrating Large Language Models, Knowledge Graphs and Driving Rules (AAAI 2025 · similarity 0.88)
- Uncertainty-Guided Never-Ending Learning to Drive (CVPR 2024 · similarity 0.87)
- Prioritizing Perception-Guided Self-Supervision: A New Paradigm for Causal Modeling in End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.87)
- Continuously Learning, Adapting, and Improving: A Dual-Process Approach to Autonomous Driving (NeurIPS 2024 · similarity 0.87)
- Drive-R1: Bridging Reasoning and Planning in VLMs for Autonomous Driving with Reinforcement Learning (AAAI 2026 · similarity 0.86)
- LaST-VLA: Thinking in Latent Spatio-Temporal Space for Vision-Language-Action in Autonomous Driving (NeurIPS 2026 · similarity 0.86)
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.
