A novel approach to generalisation is presented that is able, under certain circumstances, to guarantee the generalisation to binary-output data for which no targets have been given. The basis of the guarantee is the recognition of a persistent global minimum error solution. An empirical test for whether the guarantee holds is provided which uses a technique called target reversal. The technique employs two neural networks whose convergence using opposing targets signals validity of the guarantee.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1016/s0893-6080(01)00061-2 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Deep learning classification of MGMT status of glioblastomas using multiparametric MRI with a novel domain knowledge augmented mask fusion approach.
Sci Rep
January 2025
Department of Electrical Electronical Engineering, Yaşar University, Bornova, İzmir, Turkey.
We aimed to build a robust classifier for the MGMT methylation status of glioblastoma in multiparametric MRI. We focused on multi-habitat deep image descriptors as our basic focus. A subset of the BRATS 2021 MGMT methylation dataset containing both MGMT class labels and segmentation masks was used.
View Article and Find Full Text PDFMulti-objective design of multi-material truss lattices utilizing graph neural networks.
Sci Rep
January 2025
Advanced Manufacturing Lab, ETH Zürich, Leonhardstrasse 21, 8092, Zurich, Switzerland.
The rapid advancements in additive manufacturing (AM) across different scales and material classes have enabled the creation of architected materials with highly tailored properties. Beyond geometric flexibility, multi-material AM further expands design possibilities by combining materials with distinct characteristics. While machine learning has recently shown great potential for the fast inverse design of lattice structures, its application has largely been limited to single-material systems.
View Article and Find Full Text PDFAutomated mechanical ventilator design and analysis using neural network.
Sci Rep
January 2025
School of Computing, SASTRA Deemed University, Thanjavur, Tamil Nadu, India.
Mechanical ventilation is the process through which breathing support is provided to patients who face inconvenience during respiration. During the pandemic, many people were suffering from lung disorders, which elevated the demand for mechanical ventilators. The handling of mechanical ventilators is to be done under the assistance of trained professionals and demands the selection of ideal parameters.
View Article and Find Full Text PDFAn automatic cervical cell classification model based on improved DenseNet121.
Sci Rep
January 2025
Department of Biomedical Engineering, School of Life Science and Technology, Changchun University of Science and Technology, Changchun, 130022, China.
The cervical cell classification technique can determine the degree of cellular abnormality and pathological condition, which can help doctors to detect the risk of cervical cancer at an early stage and improve the cure and survival rates of cervical cancer patients. Addressing the issue of low accuracy in cervical cell classification, a deep convolutional neural network A2SDNet121 is proposed. A2SDNet121 takes DenseNet121 as the backbone network.
View Article and Find Full Text PDFHuman response times are governed by dual anticipatory processes with distinct neural signatures.
Commun Biol
January 2025
Department of Neuroscience, University of Pennsylvania, Philadelphia, PA, USA.
Human behavior is strongly influenced by anticipation, but the underlying neural mechanisms are poorly understood. We obtained intracranial electrocephalography (iEEG) measurements in neurosurgical patients as they performed a simple sensory-motor task with variable (short or long) foreperiod delays that affected anticipation of the cue to respond. Participants showed two forms of anticipatory response biases, distinguished by more premature false alarms (FAs) or faster response times (RTs) on long-delay trials.
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!