Dataset Information

Rapid runtime learning by curating small datasets of high-quality items obtained from memory.

ABSTRACT: We propose the "runtime learning" hypothesis which states that people quickly learn to perform unfamiliar tasks as the tasks arise by using task-relevant instances of concepts stored in memory during mental training. To make learning rapid, the hypothesis claims that only a few class instances are used, but these instances are especially valuable for training. The paper motivates the hypothesis by describing related ideas from the cognitive science and machine learning literatures. Using computer simulation, we show that deep neural networks (DNNs) can learn effectively from small, curated training sets, and that valuable training items tend to lie toward the centers of data item clusters in an abstract feature space. In a series of three behavioral experiments, we show that people can also learn effectively from small, curated training sets. Critically, we find that participant reaction times and fitted drift rates are best accounted for by the confidences of DNNs trained on small datasets of highly valuable items. We conclude that the runtime learning hypothesis is a novel conjecture about the relationship between learning and memory with the potential for explaining a wide variety of cognitive phenomena.

SUBMITTER: German JS

PROVIDER: S-EPMC10578607 | biostudies-literature | 2023 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Rapid runtime learning by curating small datasets of high-quality items obtained from memory.

German Joseph Scott JS Cui Guofeng G Xu Chenliang C Jacobs Robert A RA

PLoS computational biology 20231004 10

We propose the "runtime learning" hypothesis which states that people quickly learn to perform unfamiliar tasks as the tasks arise by using task-relevant instances of concepts stored in memory during mental training. To make learning rapid, the hypothesis claims that only a few class instances are used, but these instances are especially valuable for training. The paper motivates the hypothesis by describing related ideas from the cognitive science and machine learning literatures. Using compute ...[more]

PMID: 37792896

Similar Datasets

Project description:Fear conditioning paradigms are widely used in laboratory settings to discover treatments that enhance memory consolidation and various fear processes (extinction learning, limit return of fear) that are relevant targets of exposure-based therapies. However, traditional lab-based paradigms often use the exact same conditioned stimuli for acquisition and extinction (typically differentiated with a context manipulation), whereas the opposite is true in clinical settings, as exposure therapy rarely (if ever) uses precisely the exact same stimuli from an individual's learning history. Accordingly, this study utilized a novel three-day category-based fear conditioning protocol (that uses categories of non-repeating objects [animals and tools] as conditioned stimuli during fear conditioning and extinction) to determine if aerobic exercise enhances the consolidation of extinction learning (reduces return of fear) and memory (for items encoded during extinction) during subsequent tests of extinction recall. Participants (n=40) completed a fear acquisition (day 1), fear extinction (day 2), and extinction recall (day 3) protocol. On day 1, participants completed a fear acquisition task in which they were trained to associate a category of conditioned stimuli (CS+) with the occurrence of an unconditioned stimulus (US). On day 2, participants were administered a fear extinction procedure during which CS+ and CS- categorical stimuli were presented in absence of the occurrence of the US. After completing the task, participants were randomly assigned to either receive moderate-intensity aerobic exercise (EX) or a light-intensity control (CON) condition. On day 3, participants completed fear recall tests (during which day 1, day 2, and novel CS+ and CS- stimuli were presented). Fear responding was assessed via threat expectancy ratings and skin conductance responses (SCR). During the fear recall tests, the EX group reported significantly lower threat expectancy ratings to the CS+ and CS- and exhibited greater memory of CS+ and CS- stimuli that were previously presented during day 2. There were no significant group differences for SCR. These results suggests that administration of moderate-intensity aerobic exercise following extinction learning contributes to reduced threat expectancies during tests of fear recall and enhanced memory of items encoded during extinction.

Dataset Information

Rapid runtime learning by curating small datasets of high-quality items obtained from memory.

Publications

Rapid runtime learning by curating small datasets of high-quality items obtained from memory.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets