Incidental learning across statistically-structured input in active tasks

主动任务中统计结构输入的附带学习

• 批准号: 2420979
• 负责人: Lori Holt
• 金额: $ 82.5万
• 依托单位: University of Texas at Austin
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-10-01 至 2024-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2420979&HistoricalAwards=false
• 关键词: Incidental; learning; across; statistically; structured

The natural world is rich with patterns, and organisms learn these patterns through passive exposure. This presents a powerful and flexible means of learning about the world that does not involve explicit instruction that appears to play an important role in spoken language learning. However, not all patterns can be learned by passive exposure alone. This research project investigates how learning across patterns of experience proceeds when passive exposure is insufficient to drive learning and yet there is no explicit instruction. The prior work that this project builds on suggests that real-world statistical learning may capitalize on input regularities’ global temporal alignment with behaviorally-relevant actions and events to hasten learning. Learning across statistical regularities can be incidental, and not overtly driven by an intention to learn, while still taking place in the context of an active task that generates valuable predictions and rewarding outcomes. This perspective may be transformative in how we think about human learning of statistically-structured input in complex, naturalistic environments. Findings from this research will inform the design of learning interventions that capitalize on these learning principles to be useful for diverse communities of learners. The proposed research will advance a new research approach, empirical tests of mechanistic predictions, and complementary information from behavior, electrophysiology and functional magnetic resonance imaging to understand statistical learning under more natural circumstances involving interplay among active behavior, multimodal input, selective attention, and statistical input regularities. It pursues the twin hypotheses that (1) active engagement in a rich, environment can support statistical learning by virtue of loose temporal alignment of statistically-structured input with behaviorally-relevant actions, objects, and events and (2) that this incidental statistical learning drives the emergence of selective attention to behaviorally relevant regularities, creating a virtuous cycle that promotes later learning. The team will also conduct several outreach activities including collecting data from non-university samples via a “data-truck” and providing science of learning outreach to high school students.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.

自然界有丰富的模式，有机体通过被动暴露来学习这些模式。这提供了一种强大而灵活的学习世界的方法，而不涉及似乎在口语学习中扮演重要角色的显性教学。然而，并不是所有的模式都可以通过被动暴露来学习。这个研究项目调查了当被动暴露不足以推动学习，但没有明确的指导时，跨经验模式的学习是如何进行的。该项目之前的工作表明，真实世界的统计学习可以利用输入规则与与行为相关的行动和事件的全球时间一致性来加速学习。跨越统计规律的学习可能是偶然的，并不是公开地受到学习意图的驱动，同时仍然在产生有价值的预测和有回报的结果的积极任务的背景下进行。这一观点可能会改变我们如何看待人类在复杂的自然环境中学习统计结构输入的方式。这项研究的结果将为设计学习干预措施提供参考，这些学习干预措施将利用这些学习原则对不同的学习者社区有用。这项研究将提出一种新的研究方法，即机制预测的实证检验，以及来自行为学、电生理学和功能磁共振成像的补充信息，以理解更自然情况下的统计学习，包括主动行为、多模式输入、选择性注意和统计输入规律之间的相互作用。它遵循两个假设：(1)在丰富的环境中的积极参与可以支持统计学习，因为统计结构的输入与行为相关的行动、对象和事件在时间上松散一致；(2)这种附带的统计学习推动了对行为相关规则的选择性注意的出现，创造了一个促进后来学习的良性循环。该团队还将开展几项外展活动，包括通过“数据卡车”从非大学样本中收集数据，以及向高中生提供学习科学外展。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。