One core advantage of sparse representations is the efficient coding of complex signals using compact codes. For instance, it allows for the representation of any image as a combination of few elements drawn from a large dictionary of basis functions. In the context of the efficient processing of natural images, we propose here that such codes can be optimized by designing proper homeostatic rules between the elements of the dictionary. Indeed, a common design for unsupervised learning rules relies on a gradient descent over a cost measuring representation quality with respect to sparseness. The sparseness constraint introduces a competition which can be optimized by ensuring that each item in the dictionary is selected as often as others. We implemented this rule by introducing a gain normalization similar to what is observed in a balanced neural network. We validated this theoretical insight by challenging different sparse coding algorithms with the same learning rule but with or without homeostasis. The different sparse coding algorithms were chosen for their efficiency and generality. They include least-angle regression, orthogonal matching pursuit and basis pursuit. Simulations show that for a given homeostasis rule, gradient descent performed similarly the learning of a dataset of image patches. While the coding algorithm did not matter much, including homeostasis changed qualitatively the learned features. In particular, homeostasis results in a more homogeneous set of orientation selective filters, which is closer to what is found in the visual cortex of mammals. To further validate these results, we applied this algorithm to the optimization of a visual system embedded in an aerial robot. As a consequence, this biologically-inspired learning rule demonstrates that principles observed in neural computations can help improve real-life machine learning algorithms.

Type

Publication

- This work is a followup of Perrinet, 2010, Neural Computation
- code is available @ https://github.com/laurentperrinet/BoutinRuffierPerrinet17spars and heavily uses https://github.com/bicv/SparseHebbianLearning
- the poster (PDF) will be presented Thursday, June 8 @ SPARS, Lisbon.
- see a related work describing SDPC in: (2021).