Deconfounded Lifelong Learning for Autonomous Driving via Dynamic Knowledge Spaces

2026-03-15

AI Summary

This work addresses the challenges of catastrophic forgetting, poor cross-scenario transferability, and spurious correlations induced by confounding factors in end-to-end autonomous driving under continual learning settings. To tackle these issues, the authors propose DeLL, a novel framework that uniquely integrates front-door adjustment from causal inference with nonparametric Bayesian modeling—specifically, Dirichlet process mixture models—to construct dual dynamic knowledge spaces over explicit trajectories and implicit features, without requiring a pre-specified number of clusters. This enables adaptive knowledge expansion and disentangled, deconfounded representation learning. Furthermore, DeLL incorporates a non-autoregressive evolutionary trajectory decoder for efficient motion planning. Evaluated under the CARLA closed-loop simulation and Bench2Drive protocol, DeLL demonstrates superior adaptation to novel scenarios, enhanced driving performance, and effective retention of previously acquired knowledge, validating its effectiveness in continual learning for autonomous driving.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Search and Optimization: Learning to Search
• Multiagent Systems: Multiagent Learning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

End-to-End autonomous driving (E2E-AD) systems face challenges in lifelong learning, including catastrophic forgetting, difficulty in knowledge transfer across diverse scenarios, and spurious correlations between unobservable confounders and true driving intents. To address these issues, we propose DeLL, a Deconfounded Lifelong Learning framework that integrates a Dirichlet process mixture model (DPMM) with the front-door adjustment mechanism from causal inference. The DPMM is employed to construct two dynamic knowledge spaces: a trajectory knowledge space for clustering explicit driving behaviors and an implicit feature knowledge space for discovering latent driving abilities. Leveraging the non-parametric Bayesian nature of DPMM, our framework enables adaptive expansion and incremental updating of knowledge without predefining the number of clusters, thereby mitigating catastrophic forgetting. Meanwhile, the front-door adjustment mechanism utilizes the DPMM-derived knowledge as valid mediators to deconfound spurious correlations, such as those induced by sensor noise or environmental changes, and enhances the causal expressiveness of the learned representations. Additionally, we introduce an evolutionary trajectory decoder that enables non-autoregressive planning. To evaluate the lifelong learning performance of E2E-AD, we propose new evaluation protocols and metrics based on Bench2Drive. Extensive evaluations in the closed-loop CARLA simulator demonstrate that our framework significantly improves adaptability to new driving scenarios and overall driving performance, while effectively retaining previous acquired knowledge.

Problem

• lifelong learning
• catastrophic forgetting
• spurious correlations
• knowledge transfer
• autonomous driving

Innovation

• Deconfounded Lifelong Learning
• Dynamic Knowledge Spaces
• Dirichlet Process Mixture Model
• Front-Door Adjustment
• Non-Autoregressive Planning