One strategy to program for generalization is to vary noncritical features in teaching exemplars, thereby avoiding noncritical features from being highly correlated with reinforcement and thus gaining faulty stimulus control. In the current translational evaluation, 2 groups of adults of typical development were taught to respond to arbitrary stimuli with experimenter-defined critical and noncritical features in a matching-to-sample task. The teaching arrangement used for 1 group programmed for low correlation between noncritical features and reinforcement; the teaching arrangement used for the other group programmed for high correlation between noncritical features and reinforcement. Participants in the former group displayed (a) faster acquisition of matching, (b) less variability in correct responding, and (c) a decreased likelihood of faulty stimulus control developing during training. The results contribute towards advancing the study of stimulus control and developing an explicit technology of generalization to better serve consumers of the application of our science.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1002/jaba.760 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
CT-based whole lung radiomics nomogram to identify middle-aged and elderly COVID-19 patients at high risk of progressing to critical disease.
J Appl Clin Med Phys
November 2024
Department of Radiology, Second Affiliated Hospital of Naval Medical University, Shanghai, China.
Background: COVID-19 remains widespread and poses a threat to people's physical and mental health, especially middle-aged and elderly individuals. Early identification of COVID-19 patients at high risk of progressing to critical disease helps improve overall patient outcomes and healthcare efficiency.
Purpose: To develop a radiomics nomogram to predict the risk of newly admitted middle-aged and elderly COVID-19 patients progressing to critical disease.
Supervised machine learning on ECG features to classify sleep in non-critically ill children.
J Clin Sleep Med
September 2024
Department of Neonatal and Pediatric Intensive Care, Division of Pediatric Intensive Care, Erasmus MC Sophia Children's Hospital, Rotterdam, The Netherlands.
Article Synopsis
- The study explored the potential of supervised machine learning applied to ECG data for real-time sleep monitoring in pediatric intensive care, which is currently not available.
- Researchers analyzed polysomnography recordings from 90 non-critically ill children, developing various machine learning models to classify sleep states based on derived features from the ECG data.
- Results showed that the models achieved moderate to good accuracy, especially in classifying two and three sleep states, with the XGBoost model performing best overall, highlighting the method's promise for bedside use.
Practice Reshapes the Geometry and Dynamics of Task-tailored Representations.
bioRxiv
September 2024
Department of Cognitive and Psychological Sciences, Brown University Providence, RI, U.S.
Article Synopsis
- Extensive practice improves task performance by leading to automaticity, but there are differing theories on which aspects of task representation (like low-level features or high-level memories) are affected by practice.
- Using recent developments in neural population dynamics, the study hypothesized that practice refines neural representations by optimizing their geometry to focus on the most critical task features while discarding irrelevant ones.
- Experimenting with 40 participants over 3 days, results showed that as performance improved, the highest-level context-specific representations became stronger and more stable, which directly correlated with better task efficiency and reduced errors when switching tasks.
Laboratory data and broncho-alveolar lavage on Covid-19 patients with no intensive care unit admission: Correlation with chest CT features and clinical outcomes.
Medicine (Baltimore)
July 2024
Department of Experimental and Clinical Biomedical Sciences, Radiodiagnostic Unit n. 2, University of Florence - Azienda Ospedaliero-Universitaria Careggi, Florence, Italy.
Broncho-alveolar lavage (BAL) is indicated in cases of uncertain diagnosis but high suspicion of Sars-Cov-2 infection allowing to collect material for microbiological culture to define the presence of coinfection or super-infection. This prospective study investigated the correlation between chest computed tomography (CT) findings, Covid-19 Reporting and Data System score, and clinical outcomes in Coronavirus disease 2019 (Covid-19) patients who underwent BAL with the aim of predicting outcomes such as lung coinfection, respiratory failure, and hospitalization length based on chest CT abnormalities. Study population included 34 patients (range 38-90 years old; 20 males, 14 females) with a positive nucleic acid amplification test for Covid-19 infection, suitable BAL examination, and good quality chest CT scan in the absence of lung cancer history.
View Article and Find Full Text PDFFDAA: A feature distribution-aware transferable adversarial attack method.
Neural Netw
October 2024
School of Computer Science and Engineering, South China University of Technology, Guangzhou 510006, Guangdong, China.
In recent years, the research on transferable feature-level adversarial attack has become a hot spot due to attacking unknown deep neural networks successfully. But the following problems limit its transferability. Existing feature disruption methods often focus on computing feature weights precisely, while overlooking the noise influence of feature maps, which results in disturbing non-critical features.
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!