Luo Xiaoliang, Roads Brett D, Love Bradley C
Department of Experimental Psychology, University College London, 26 Bedford Way, London, WC1H 0AP UK.
Comput Brain Behav. 2021;4(2):213-230. doi: 10.1007/s42113-021-00098-y. Epub 2021 Feb 12.
People deploy top-down, goal-directed attention to accomplish tasks, such as finding lost keys. By tuning the visual system to relevant information sources, object recognition can become more efficient (a benefit) and more biased toward the target (a potential cost). Motivated by selective attention in categorisation models, we developed a goal-directed attention mechanism that can process naturalistic (photographic) stimuli. Our attention mechanism can be incorporated into any existing deep convolutional neural networks (DCNNs). The processing stages in DCNNs have been related to ventral visual stream. In that light, our attentional mechanism incorporates top-down influences from prefrontal cortex (PFC) to support goal-directed behaviour. Akin to how attention weights in categorisation models warp representational spaces, we introduce a layer of attention weights to the mid-level of a DCNN that amplify or attenuate activity to further a goal. We evaluated the attentional mechanism using photographic stimuli, varying the attentional target. We found that increasing goal-directed attention has benefits (increasing hit rates) and costs (increasing false alarm rates). At a moderate level, attention improves sensitivity (i.e. increases ) at only a moderate increase in bias for tasks involving standard images, blended images and natural adversarial images chosen to fool DCNNs. These results suggest that goal-directed attention can reconfigure general-purpose DCNNs to better suit the current task goal, much like PFC modulates activity along the ventral stream. In addition to being more parsimonious and brain consistent, the mid-level attention approach performed better than a standard machine learning approach for transfer learning, namely retraining the final network layer to accommodate the new task.
人们运用自上而下的、目标导向的注意力来完成任务，比如寻找丢失的钥匙。通过将视觉系统调整到相关信息源，物体识别可以变得更高效（一个好处），同时对目标的偏向性更强（一个潜在代价）。受分类模型中选择性注意力的启发，我们开发了一种能够处理自然主义（照片）刺激的目标导向注意力机制。我们的注意力机制可以被整合到任何现有的深度卷积神经网络（DCNN）中。DCNN中的处理阶段与腹侧视觉通路有关。鉴于此，我们的注意力机制纳入了来自前额叶皮层（PFC）的自上而下的影响，以支持目标导向行为。类似于分类模型中的注意力权重如何扭曲表征空间，我们在DCNN的中间层引入了一层注意力权重，以增强或减弱活动来推进目标。我们使用照片刺激评估了注意力机制，并改变了注意力目标。我们发现增加目标导向注意力有好处（提高命中率）也有代价（提高误报率）。在适度水平上，对于涉及标准图像、混合图像和用于欺骗DCNN的自然对抗图像的任务，注意力在仅适度增加偏向性的情况下提高了敏感度（即增加了 ）。这些结果表明，目标导向注意力可以重新配置通用的DCNN，使其更好地适应当前任务目标，就像PFC调节腹侧通路的活动一样。除了更简洁且与大脑一致外，中间层注意力方法在迁移学习方面比标准机器学习方法（即重新训练最终网络层以适应新任务）表现更好。
