Researchers unveil a new 'Centered Daydreaming' algorithm enabling AI to effectively learn and recall from imbalanced, real-world image data.
Key Details
- 1The Daydreaming algorithm, inspired by sleep-memory consolidation in biological brains, improves Hopfield network storage capacity up to the theoretical maximum.
- 2Original version struggled with heavily biased or unbalanced real-world image data (e.g., mostly white or black pixels).
- 3The new 'Centered Daydreaming' approach focuses on local differences rather than absolute values, enhancing performance with realistic data.
- 4Published in the Journal of Statistical Mechanics: Theory and Experiment (July 15, 2026).
- 5The update maintains effective associative memory performance even when data distributions are heavily skewed.
Why It Matters
Handling biased or imbalanced image data is a core challenge in applying neural networks to medical imaging, where real scans often deviate from ideal distributions. Biologically inspired improvements like this could increase reliability, interpretability, and efficiency of future AI models for radiology.
Source
EurekAlert
Related News
•EurekAlert
UVA's $4M NIH Grant Backs AI to Optimize Cardiac MRI Therapies
UVA secures $4 million NIH funding to develop GenStrain, an AI model using MRI data to personalize cardiac pacing therapies.
•EurekAlert
Machine Learning Models Predict Cancer Risk with Routine Health Data
Researchers developed machine learning models using routinely collected health data to predict individual risk for 20 common cancers, outperforming traditional methods for several cancer types.
•EurekAlert
Duke Receives $25M to Lead National Trustworthy AI Hub for Dementia and Aging
Duke University has launched a $25 million national center to make AI for aging and dementia clinically robust and trustworthy.