权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

CAREER: Contingent Learning in a Structured World

职业：结构化世界中的条件学习

基本信息

批准号：
1846578
负责人：
Erie Boorman
金额：
$ 75.39万
依托单位：
University of California-Davis
依托单位国家：
美国
项目类别：
Standard Grant
财政年份：
2019
资助国家：
美国
起止时间：
2019-08-01 至 2024-07-31
项目状态：
已结题

来源：
https://www.nsf.gov/awardsearch/showAward?AWD_ID=1846578&HistoricalAwards=false
关键词：
CAREER Contingent Learning Structured World

项目摘要

How the brain forms, tunes, and uses predictive models that specify the causal links between stimuli in the environment, our choices, and their outcomes is a fundamental question in Psychology and Neuroscience. In order to make good predictions in a complex world, the brain needs to attribute good and bad outcomes to their most likely causes, a problem known as "credit assignment". There is presently limited understanding of how people attribute outcomes to their most likely causes in real-world environments. Recent evidence suggests that humans can use an understanding of the environment's causal structure to attribute outcomes to their most likely causes ("model-based credit assignment"), rather than only attributing them to the most recently made choices ("model-free credit assignment"), as standard models posit. Dr. Boorman's research will develop the first neurally-inspired theory of model-based credit assignment. The insights from this basic science research have the potential to inform (1) theories about human cognition more broadly; (2) targeted education programs that leverages a better understanding of how learning works in the human brain; (3) mechanistic predictions for application to Artificial Intelligence research; and (4) principled investigations into potential biomarkers and treatment targets for psychiatric disorders such as Schizophrenia and PTSD, with deficits that include abnormal credit assignment.The overarching goal of this proposal is to elucidate how the human brain attributes reinforcement outcomes to their likely causes, or "credit assignment". Recent evidence suggests the brain can harness a model of the environment or task structure to assign credit for outcomes adaptively ("model-based credit assignment"), in addition to the most recently made choices ("model-free credit assignment"). Dr. Boorman's proposed research will integrate multivariate analysis methods with Bayesian models that formalize updating and inference, and apply them to fMRI, EEG, and TMS data. Insights from this fundamental research will inform mechanistic models of human learning, theories of cognition more broadly, reinforcement learning in Artificial Intelligence research, and investigations into psychiatric disorders with deficits that include abnormal credit assignment, such as Schizophrenia and PTSD.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

大脑如何形成、调整和使用预测模型，这些模型指定了环境中的刺激、我们的选择及其结果之间的因果关系，这是心理学和神经科学中的一个基本问题。为了在复杂的世界中做出正确的预测，大脑需要将好的和坏的结果归因于最可能的原因，这就是所谓的“信用分配”问题。目前，人们对现实世界环境中人们如何将结果归因于最可能的原因的理解有限。最近的证据表明，人类可以利用对环境因果结构的理解来将结果归因于最可能的原因（“基于模型的信用分配”），而不是像标准模型那样仅将其归因于最近做出的选择（“无模型信用分配”）。 Boorman博士的研究将开发第一个基于模型的信用分配神经启发理论。这项基础科学研究的见解有可能为以下方面提供信息：（1）更广泛的人类认知理论;（2）有针对性的教育计划，以更好地了解人类大脑中的学习方式;（3）应用于人工智能研究的机械预测;和（4）对精神障碍如精神分裂症和创伤后应激障碍的潜在生物标志物和治疗靶点的原则性研究，与赤字，包括异常信用分配。这一建议的首要目标是阐明人脑如何属性强化结果，以他们可能的原因，或“信用分配”。最近的证据表明，大脑可以利用环境或任务结构的模型来自适应地为结果分配信用（“基于模型的信用分配”），以及最近做出的选择（“无模型信用分配”）。Boorman博士提出的研究将整合多变量分析方法与贝叶斯模型，形式化更新和推理，并将其应用于fMRI，EEG和TMS数据。这项基础研究的见解将为人类学习的机械模型、更广泛的认知理论、人工智能研究中的强化学习以及对包括异常信用分配在内的精神疾病的调查提供信息，该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的智力价值和更广泛的评估来支持。影响审查标准。