Decision prioritization and causal reasoning in decision hierarchies.
Decision prioritization and causal reasoning in decision hierarchies.
复制标题
DOI：
10.1371/journal.pcbi.1009688
复制
发表时间：
2021-12
影响因子：
4.3
通讯作者：
Zylberberg A
中科院分区：
文献类型：
--
作者：
Zylberberg A
从做饭到找到一条到达目的地的路线，许多真实的生活决策都可以分解为一系列子决策。在层次结构中，选择要考虑的决策需要在可能的决策序列的潜在巨大空间上进行规划。为了深入了解人们如何决定决定什么，我们研究了一个结合感知决策，主动感知和分层和反事实推理的新任务。人类参与者必须找到隐藏在决策树最底层的目标。他们可以从决策树的不同节点请求信息，以收集有关目标位置的噪声证据。只有在叶节点出现错误后才给出反馈，并提供了关于错误原因的模糊证据。尽管任务的复杂性（有107个潜在状态），参与者能够有效地计划任务。这一过程的计算模型确定了少量计算复杂度低的解释人类行为的方法。这些策略包括在决策树的分支点做出分类决策，而不是向前推进整个概率分布，丢弃被认为不可靠的感官证据来做出选择，并在初始计划失败后使用选择信心来推断错误的原因。基于概率推理或短视抽样规范的计划无法捕捉参与者的行为。我们的研究结果表明，它是可能的，以确定标志性的启发式规划与人类行为中的传感器和使用任务的中间复杂性有助于确定规则的基础上的人类能力的决策层次结构的原因。复杂的决策通常被分解为一系列信息收集行动，然后是寻求奖励的行动。例如，医生可以在建议纠正措施之前进行一系列测试以诊断患者的疾病。人们如何决定接下来要问的问题（测试、实验、查询）是什么？人类参与者被展示了一个分叉三次的二叉决策树。他们可以从分叉点收集信息，以收集关于目标位置的噪音证据。我们确定了人们在这项复杂任务中用于有效计划的策略。参与者利用任务的层次结构，并依赖于对过去决策的信心来选择后续行动。我们的研究结果对人们如何在大的部分可观察的领域有效地规划产生了影响，并对人工代理的设计产生了影响，这些代理必须通过积极的探索做出决定，并对人类和其他动物的规划进行神经生理学研究。
From cooking a meal to finding a route to a destination, many real life decisions can be decomposed into a hierarchy of sub-decisions. In a hierarchy, choosing which decision to think about requires planning over a potentially vast space of possible decision sequences. To gain insight into how people decide what to decide on, we studied a novel task that combines perceptual decision making, active sensing and hierarchical and counterfactual reasoning. Human participants had to find a target hidden at the lowest level of a decision tree. They could solicit information from the different nodes of the decision tree to gather noisy evidence about the target’s location. Feedback was given only after errors at the leaf nodes and provided ambiguous evidence about the cause of the error. Despite the complexity of task (with 107 latent states) participants were able to plan efficiently in the task. A computational model of this process identified a small number of heuristics of low computational complexity that accounted for human behavior. These heuristics include making categorical decisions at the branching points of the decision tree rather than carrying forward entire probability distributions, discarding sensory evidence deemed unreliable to make a choice, and using choice confidence to infer the cause of the error after an initial plan failed. Plans based on probabilistic inference or myopic sampling norms could not capture participants’ behavior. Our results show that it is possible to identify hallmarks of heuristic planning with sensing in human behavior and that the use of tasks of intermediate complexity helps identify the rules underlying human ability to reason over decision hierarchies. Complex decisions are often broken down into a sequence of information-gathering actions followed by reward-seeking actions. For example, a physician may conduct a series of tests to diagnose a patient’s disease before suggesting a corrective action. How do people decide what is the appropriate question (test, experiment, query) to ask next? Human participants were presented with a binary decision tree that bifurcated three times. They could solicit information from the bifurcation points to gather noisy evidence about the location of a target. We identified the heuristics that people used to plan efficiently in this complex task. Participants exploited the hierarchical structure of the task and relied on the confidence in past decision to inform the selection of subsequent actions. Our results bear on how people plan efficiently in large partially observable domains, and have implications for the design of artificial agents that have to make decisions with active exploration and for neurophysiological studies of planning in humans and other animals.
登录
查看更多内容
影响因子：
3.4
作者：
Ais, Joaquin;Zylberberg, Ariel;Sigman, Mariano
通讯作者：
Sigman, Mariano
影响因子：
16.2
作者：
Daw ND;Gershman SJ;Seymour B;Dayan P;Dolan RJ
通讯作者：
Dolan RJ
DOI：
10.1073/pnas.1710913114
发表时间：
2017-10-03
影响因子：
11.1
作者：
Calderon, Cristian Buc;Dewulf, Myrtille;Verguts, Tom
通讯作者：
Verguts, Tom
影响因子：
16.2
作者：
Balaguer J;Spiers H;Hassabis D;Summerfield C
通讯作者：
Summerfield C
影响因子：
4.3
作者：
Adler, William T.;Ma, Wei Ji
通讯作者：
Ma, Wei Ji