This work breaths from the paper from nature to develop Artificial Intelligence advacement through brain reverse engineering and bio inspiration. You can find the paper here http://www.cell.com/neuron/pdf/S0896-6273(17)30509-3.pdf
There is a part coming from waitbutwhy neurallink article that explores the functions behind the limbic system, aiming for it as an advanced survival system . https://waitbutwhy.com/2017/04/neuralink.html
Information about Neural Net architectures http://www.deeplearningbook.org
The limbic system is a set of brain structures located on both sides of the thalamus, immediately beneath the cerebrum.[1] It has also been referred to as the paleomammalian cortex. It is not a separate system but a collection of structures from the telencephalon, diencephalon, and mesencephalon.[2] It includes the olfactory bulbs, hippocampus, hypothalamus, amygdala, anterior thalamic nuclei, fornix, columns of fornix, mammillary body, septum pellucidum, habenular commissure, cingulate gyrus, parahippocampal gyrus, limbic cortex, and limbic midbrain areas.[3]
The limbic system supports a variety of functions including emotion, behavior, motivation, long-term memory, and olfaction.[4] Emotional life is largely housed in the limbic system, and it has a great deal to do with the formation of memories.
In this repository you can find several experiments to reverse eingeniering documentation Data gathering. Cognitive Experiments. Neural Net Experimentation , from classification to Architecture model Design
Find bellow a descripcion of main parts and ideas regarding neural nets that could be implemented
THE AMYGDALA : Responds to fear, danger and anxiety. There are two parts of the amygdala that show an adversarial behaviour Type of neural net that could be implemented : perceptron, adversarial neural net, logistic regression. As other ideas, one shot learning -as a quick response to fear-
THE HIPPOCAMPUS : memory survival mode approach Type of neural net that could be implemented : reinforced LSTM,
THE THALAMUS : sensory middle-dispositive conected with the cortex Type of neural net that could be implemented : CNNs, RNNs, InceptionV3 and InceptionV4
ORBITOFRONTAL CORTEX :image sensatory input
SENSORY CORTEX : The sensory cortex process the senses information - visual, olfatory - and separates the sensory data
As inputs, find two kinds differenciated: SI - Sensory input. Images and words ES - Emotional signal
Please, note this is a work under development
You specially can contribute in NeuralNet design and experimentation, and / or correcting any misinterpretation or incorrect concept that you might find . You can also recommend some lectures or papers that may be of interest
- Datasets
- Datasets
- Thalamus structure
- Discovery PoCs
- Inspiration links
- Neural Net description
- Architectural design