Deans' stroke musings: Attention-Gated Brain Propagation: How the brain can implement reward-based error backpropagation

Changing stroke rehab and research worldwide now.Time is Brain! trillions and trillions of neurons that DIE each day because there are NO effective hyperacute therapies besides tPA(only 12% effective). I have 523 posts on hyperacute therapy, enough for researchers to spend decades proving them out. These are my personal ideas and blog on stroke rehabilitation and stroke research. Do not attempt any of these without checking with your medical provider. Unless you join me in agitating, when you need these therapies they won't be there.

What this blog is for:

My blog is not to help survivors recover, it is to have the 10 million yearly stroke survivors light fires underneath their doctors, stroke hospitals and stroke researchers to get stroke solved. 100% recovery. The stroke medical world is completely failing at that goal, they don't even have it as a goal. Shortly after getting out of the hospital and getting NO information on the process or protocols of stroke rehabilitation and recovery I started searching on the internet and found that no other survivor received useful information. This is an attempt to cover all stroke rehabilitation information that should be readily available to survivors so they can talk with informed knowledge to their medical staff. It lays out what needs to be done to get stroke survivors closer to 100% recovery. It's quite disgusting that this information is not available from every stroke association and doctors group.

Monday, November 9, 2020

Attention-Gated Brain Propagation: How the brain can implement reward-based error backpropagation

Way beyond my ability to decipher if this has any applicability to stroke rehab. So ask your doctor for guidance on this.

Attention-Gated Brain Propagation: How the brain can implement reward-based error backpropagation

Isabella Pozzi Machine Learning Group Centrum Wiskunde & Informatica Amsterdam, The Netherlands isabella.pozzi@cwi.nl

Sander M. Bohté Machine Learning Group Centrum Wiskunde & Informatica Amsterdam, The Netherlands s.m.bohte@cwi.nl

Pieter R. Roelfsema Vision & Cognition Group Netherlands Institute for Neuroscience Amsterdam, The Netherlands p.roelfsema@nin.knaw.nl

Abstract

Much recent work has focused on biologically plausible variants of supervised learning algorithms. However, there is no teacher in the motor cortex that instructs the motor neurons and learning in the brain depends on reward and punishment. We demonstrate a biologically plausible reinforcement learning scheme for deep networks with an arbitrary number of layers. The network chooses an action by selecting a unit in the output layer and uses feedback connections to assign credit to the units in successively lower layers that are responsible for this action. After the choice, the network receives reinforcement and there is no teacher correcting the errors. We show how the new learning scheme – Attention-Gated Brain Propagation (BrainProp) – is mathematically equivalent to error backpropagation, for one output unit at a time. We demonstrate successful learning of deep fully connected, convolutional and locally connected networks on classical and hard image-classification benchmarks; MNIST, CIFAR10, CIFAR100 and Tiny ImageNet. BrainProp achieves an accuracy that is equivalent to that of standard error-backpropagation, and better than state-of-the-art biologically inspired learning schemes. Additionally, the trial-and-error nature of learning is associated with limited additional training time so that BrainProp is a factor of 1-3.5 times slower. Our results thereby provide new insights into how deep learning may be implemented in the brain.