Neural Networks in Simbrain

Neural Networks in Simbrain
https://www.youtube.com/watch?v=3jYe0ruxtOs

* This video contains audio only as music:
"Transcend" by Audionautix is licensed under a Creative Commons Attribution license (https://creativecommons.org/licenses/...)
Artist: http://audionautix.com/
Genre: Ambient; Mood: Calm ; Type: Instrumental.


You can see that if the input pattern of neural activations (set of initial potential values in the multi-parametric network) changes – the overall value of the network still converges towards a common point (attractor network). This is because the neurons and connections have not been modified. When the system is changed (ie. synaptic connection weights are modified (therefore affecting the internodal learning algorithms) or when neurons degenerate & are lost in disease), then the local mimimum in the search space landscape changes (new attractor point).


I am interested in modelling various systems in the brain:
--- MTL/Hippocampal/CA3 neural networks. Hopfield recurrent networks, attractor network, with pattern converging towards a common pattern (memory), based on initial cues. Auto-associative networks in CA3, content-addressable (not location-addressable) memory – re-construction of whole memory from one or more cues. Sparse code and system of overlaps – memories shared info (are associated) but non-spurious distinct memories can be formed from the system. Pattern separation and auto-association and pattern completion processes. Unsupervised self-organising topographic maps, STDP, rate coding, and oscillatory interference, for representation of allocentric space, context-based episodic memory, predictive navigation of scenes.
--- V1-V4/V5 system of deep convolutional neural networks for visual feature representations.
--- Limbic/Amygdala/OFC system for reward-based and aversion-based semi-supervised reinforcement learning.
--- The PFC neural networks for top-down conscious decision-making.
--- Somatosensory and somatomotor systems.
see https://www.bioneurotech.com




SimBrain:

http://simbrain.net/
https://github.com/simbrain

Tosi Z, Yoshimi J (2016) “Simbrain 3.0: A flexible, visually-oriented neural network simulator”. Journal of Neural Networks. PMID: 27541049. doi: https://doi.org/10.1016/j.neunet.2016...



Tensorflow Playground:

https://playground.tensorflow.org/




I used EZ-vid for screen capture and Filmora for video-editing. I would strongly recommended these to anyone else interested in making videos:

https://www.ezvid.com/
https://filmora.wondershare.com/
Category
Science & Technology