Transparent Machine Learning Using Brain-Motivated Optimization
Wednesday, January 11, 2017 6:30 PM to 9:00 PM Action Spot Startup Place 99 Wilson Ave, San Jose, CA , San Jose, CA (map) Price: $5.00 /per person Refund policy Feed-forward networks have been the basis of Artificial Neural Networks, Deep Convolution, and Recurrent Networks. However the internal decision processes of feed-forward networks are difficult to explain: they are known to be a “black-box”. For decades Neuroscientists have argued that brain networks are not feed-forward. We have developed a new type of neural network which is not feed-forward. It performs optimization during recognition – not to learn weights, but to determine neuron activation. This allows learning to be simpler, the network to be more updatable, and the internal decision process easier to understand. Moreover we are now able to convert feedforward networks to our form and explain…
Link to Full Article: Transparent Machine Learning Using Brain-Motivated Optimization
Sharing
