arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2005.04298cs.CVcs.LGcs.RO
收藏

Attentional Bottleneck:迈向可解释的深度驱动网络

Attentional Bottleneck: Towards an Interpretable Deep Driving Network

Waymo Research(Waymo研究院)

Jinkyu Kim, Mayank Bansal

摘要:

深度神经网络是自动驾驶汽车行为预测和运动生成的关键组成部分。它们的主要缺点之一是缺乏透明度:它们应该为触发某些行为的原因提供易于解释的理由。我们提出了一个架构称为注意力瓶颈的目标是提高透明度。我们的关键思想是将视觉注意力(识别模型正在使用输入的哪些方面)与信息瓶颈(使模型能够只使用输入的重要方面)相结合。这不仅提供了稀疏和可解释的注意力地图(例如,只关注场景中的特定车辆),而且它在不影响模型准确性的情况下增加了这种透明度。事实上,我们发现在将注意力瓶颈应用于ChauffeurNet模型时,准确性略有提高,而我们发现传统视觉注意力模型的准确性有所下降。

英文摘要:

Deep neural networks are a key component of behavior prediction and motion generation for self-driving cars. One of their main drawbacks is a lack of transparency: they should provide easy to interpret rationales for what triggers certain behaviors. We propose an architecture called Attentional Bottleneck with the goal of improving transparency. Our key idea is to combine visual attention, which identifies what aspects of the input the model is using, with an information bottleneck that enables the model to only use aspects of the input which are important. This not only provides sparse and interpretable attention maps (e.g. focusing only on specific vehicles in the scene), but it adds this transparency at no cost to model accuracy. In fact, we find slight improvements in accuracy when applying Attentional Bottleneck to the ChauffeurNet model, whereas we find that the accuracy deteriorates with a traditional visual attention model.

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。

↑