ETA：通过前瞻思考实现效率，一种使用大模型的自动驾驶双系统方法
ETA: Efficiency through Thinking Ahead, A Dual Approach to Self-Driving with Large Models
AI总结：
针对自动驾驶中大模型推理延迟问题，本文提出异步双系统 ETA，通过前瞻批量推理、特征传播与动作掩码融合，在 Bench2Drive 上以 50 ms 速度取得 69.53 分。
摘要：
我们如何才能在不牺牲推理速度的情况下受益于大模型？这是自动驾驶系统中的一个常见困境。一种流行的解决方案是双系统架构，即使用小模型进行快速的反应式决策，并使用较大的模型进行较慢但信息量更大的分析。现有双系统设计通常采用并行架构：要么在每个当前帧直接使用大模型进行推理，要么从先前存储的推理结果中检索。然而，这些工作仍难以让大模型对每个在线帧作出及时响应。我们的关键洞见是，将当前帧的密集计算转移到先前的时间步，并对多个时间步执行批量推理，从而使大模型能够迅速响应每个时间步。为实现这种转移，我们提出了 Efficiency through Thinking Ahead（ETA，通过前瞻思考实现效率），这是一个异步系统，旨在：（1）利用大模型的未来预测，将来自过去的信息丰富特征传播到当前帧；（2）使用小模型提取当前帧特征，以实现实时响应；（3）通过动作掩码机制融合这些双系统特征，该机制会强调对动作至关重要的图像区域。在 Bench2Drive CARLA Leaderboard-v2 基准上进行评估时，ETA 将最先进性能提升了 8%，驾驶分数达到 69.53，同时保持 50 ms 的近实时推理速度。
英文摘要：
How can we benefit from large models without sacrificing inference speed, a common dilemma in self-driving systems? A prevalent solution is a dual-system architecture, employing a small model for rapid, reactive decisions and a larger model for slower but more informative analyses. Existing dual-system designs often implement parallel architectures where inference is either directly conducted using the large model at each current frame or retrieved from previously stored inference results. However, these works still struggle to enable large models for a timely response to every online frame. Our key insight is to shift intensive computations of the current frame to previous time steps and perform a batch inference of multiple time steps to make large models respond promptly to each time step. To achieve the shifting, we introduce Efficiency through Thinking Ahead (ETA), an asynchronous system designed to: (1) propagate informative features from the past to the current frame using future predictions from the large model, (2) extract current frame features using a small model for real-time responsiveness, and (3) integrate these dual features via an action mask mechanism that emphasizes action-critical image regions. Evaluated on the Bench2Drive CARLA Leaderboard-v2 benchmark, ETA advances state-of-the-art performance by 8% with a driving score of 69.53 while maintaining a near-real-time inference speed at 50 ms.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
