Jay McClelland
(Stanford University)
Welcome and Update
Date: February 26, 2014
Suruya Ganguli
(Stanford University)
Introduction and Overview
Date: February 26, 2014
Bruno Olshausen
(Berkeley)
Sparse coding in brains and machines
Date: February 26, 2014
Description:
The principle of sparse coding figures prominently in both theories of brain function and in the design of signal processing and computer vision systems. Here I will review what we know about sparse coding in brains, and some of the pressing questions that could be resolved by the new recording technologies now emerging. I will then turn to what we have learned from technological applications about the utility of sparse coding, and what this may tell us about neural representation. In particular, the emerging paradigm of ‘deep networks’ offers an opportunity for interdisciplinary exploration into theories of neural coding, where insights gained from one field of inquiry may inform the other.
David Donoho
(Stanford University)
Sparsity and Optimization -- some history and some questions
Date: February 26, 2014
Description:
Sparsity-seeking and sparsity-exploiting algorithms have been very popular in signal and image processing in recent years. I will review some of the earlier examples, such as blind deconvolution and superresolution, and transition to more recent ones such as dictionary learning and compressed sensing. Olshausen and Field had a landmark paper in the mid 90’s connecting one such sparsity-seeking algorithm and vision. And after Barlow, it seems that sparsity-seeking algorithms would make a very good strategy for processing natural stimuli. In general, it seems that the kind of nonlinear optimization algorithms that make for interesting ‘math papers’ do not connect with biological ideas. I hope to provoke discussion.
Dmitri Chklovskii
(Janelia Farm, HHMI)
Towards a holistic view of single-neuron computation
Date: February 26, 2014
Description:
A neuron is a basic physiological and computational unit of the brain. While much is known about the physiological properties of a neuron, its computational role is poorly understood. Here we propose to view a neuron as a signal processing device that represents the incoming streaming data matrix as a sparse vector of synaptic weights scaled by an outgoing sparse activity vector… Our theoretical framework makes several predictions, some of which can be verified by the existing data, others require further experiments. Such framework should allow modeling the function of neuronal circuits without necessarily measuring all the microscopic biophysical parameters, as well as facilitate the design of neuromorphic electronics.
Surya Ganguli
(Stanford University)
The Role of Sparsity in Neural Computation and the Natural World: Panel Discussion
Date: February 26, 2014
Symposium Poster
