Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..

Embodied Cognition Augmented End2End Autonomous Driving

Authors: Ling Niu, Xiaoji Zheng, han wang, Ziyuan Yang, Chen Zheng, Bokui Chen, Jiangtao Gong

NeurIPS 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Both open-loop and closed-loop tests are conducted for a comprehensive evaluation of planning performance. Experimental results demonstrate that the E3AD paradigm significantly enhances the end-to-end planning performance of baseline models. Ablation studies further validate the contribution of driving cognition and the effectiveness of comparative learning process.
Researcher Affiliation Academia Ling Niu Tsinghua University EMAIL Xiaoji Zheng Tsinghua University EMAIL Han Wang Tsinghua University EMAIL Chen Zheng Tsinghua University EMAIL Ziyuan Yang Tsinghua University EMAIL Bokui Chen* Tsinghua University EMAIL Jiangtao Gong Tsinghua University EMAIL
Pseudocode No The paper describes methods in text and equations, such as in Section 3.2, 3.3, 3.4, 3.5, and 3.6, and presents a training pipeline in Figure 1, but does not include any clearly labeled pseudocode or algorithm blocks.
Open Source Code No Our code will be made available at https://github.com/AIR-DISCOVER/E-cubed-AD.
Open Datasets Yes We conducted experiments on the publicly available nu Scenes dataset [1], which provides 1,000 urban driving scenes under diverse weather conditions. [1] Holger Caesar, Varun Bankiti, Alex H Lang, Sourabh Vora, Venice Erin Liong, Qiang Xu, Anush Krishnan, Yu Pan, Giancarlo Baldan, and Oscar Beijbom. nuscenes: A multimodal dataset for autonomous driving. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11621 11631, 2020.
Dataset Splits Yes We split the dataset into training set validation, and test set. The structured data was shuffled and divided into two parts with a ratio of 80:10:10. The training set comprised 1894 clips (1037 from expert drivers and 857 from novice drivers), the validation set consists of 236 clips (144 expert, and 92 novice), and the test set contained 237 clips (118 from expert drivers and 119 from novice drivers). For nuScenes: The dataset includes 700 training, 150 validation, and 150 test scenes, each lasting about 20 s with keyframes annotated at 2 Hz.
Hardware Specification Yes Training on the full dataset requires approximately 12 h on a single NVIDIA A100 40 GB GPU.
Software Dependencies Yes Data were recorded in real time via CURRY 9
Experiment Setup Yes We train the model on aligned 2s video EEG segment pairs with a batch size of 16 for 120 epochs using Adam (learning rate 2e-5 on both backbones and adapters, lr ratio 1:1), dropout of 0.01, and weight decay of 1e-5.