郑重声明:原文参见标题,如有侵权,请联系作者,将会撤销发布!
2. Reward-Prediction Error Meets Dopamine
3. Reward-Prediction Error and Incentive Salience: What Do They Explain?
4. Explanatory Depth, Reward-Prediction Error and Incentive Salience
4.1. Depth as scope, reward-prediction error and incentive salience
4.2. Depth as invariance, reward-prediction error and incentive salience
根据多巴胺的奖励预测误差假设(RPEH),中脑多巴胺能神经元的相位活动表示特定事件的预测奖励与当前经历的奖励之间存在差异。可以说这个假设是深刻,优雅和美丽的,代表了计算神经科学的最大成功之一。本文研究了这种说法,为现有文献做出了两点贡献。首先,它对公式化定义RPEH和随后获得成功的主要步骤进行了全面的历史描述。其次,根据这一历史记录,它解释了RPEH在哪种意义上具有解释性,在何种情况下可以合理地认为它比多巴胺的刺激显著性假设更深远,多巴胺可以说是目前RPEH最重要的替代方案。
Keywords: 多巴胺(Dopamine);奖励预测误差(Reward-Prediction Error);解释深度(Explanatory Depth);刺激显著性(Incentive Salience);强化学习(Reinforcement Learning)
2. Reward-Prediction Error Meets Dopamine
3. Reward-Prediction Error and Incentive Salience: What Do They Explain?
4. Explanatory Depth, Reward-Prediction Error and Incentive Salience
4.1. Depth as scope, reward-prediction error and incentive salience
4.2. Depth as invariance, reward-prediction error and incentive salience