sparsecoding
nounUK /spɑːs ˈkəʊdɪŋ/
US /spɑːrs ˈkoʊdɪŋ/
Sparse coding is a way of storing information using only a few active parts at a time.
The brain uses sparse coding so that only a small group of neurons fires for each memory.
OTHER MEANINGS
2. A machine learning method where most values in a data representation are zero.
The model learned a sparse coding of the image, keeping only the most important features.
3. A technique in signal processing that represents a signal with few nonzero coefficients.
Sparse coding helped compress the audio file without losing much quality.
4. A theory in neuroscience suggesting that only a small fraction of neurons respond to any given stimulus.
Studies of the visual cortex support sparse coding during object recognition.