Computational Neuroscience Flashcards
What kinds of neuroscience can be classified as computational neuroscience?
Any type of neuroscience that employs mathematical models, computer simulations, or theoretical analysis
The idea of computational neuroscience started in _____?
in the 1980s
How has today’s computational neuroscience field grown from then?
there’s more computing power to do realistic simulations of neural systems and new insights are being made about the study of simplified model individual neurons to large large networks of them
What is computational neuronal modeling?
this is the use of computer algorithms or simulator programs to model the behavior of neurons and learn more about their process.
What can computational neuronal modeling be used for?
This can be used for studying how neurons process,transmit info, how they react to damage or disease, explore the underlying mechanisms of neural networks
Give two examples of computational neuronal modeling?
Biophysical modeling and Neural Network modeling
Explain what each of the examples aims to model?
Biophysical models - recreate the detailed behavior of individual neurons
Neural Network Models - simulate the interactions between multiple neuron groups in a large group of cells and large scale models can simulate the activity of entire brain regions or subsystems within the nervous system.
Why can we not make a true AI with the knowledge of neuroscience that we have?
We don’t know enough about how cognition and complex thinking come from the firing of neurons
What else would we need to know in order to build an AI?
1) understand what makes a certain human mind unique
2) what drives imagination or creativity
3) how the subconscious mind works
4) which synapses support each of these functions
How has AI improved over the years and what can it do now (give some examples)?
imitating how the brain performs certain well-
understood computations.
What are artificial neural networks?
These are created to mimic existing networks of neurons in the brain
How do they process information, and how is it similar to how our neurons process stuff?
These artificial networks consist of layers of nodes that are analogous to neurons. The nodes are connected to
each other by a series of mathematical weights that act like the synapses between neurons, each with a particular
strength that can be changed based on experience, just like real synapses. Data are fed into the layers of nodes, which then multiplies those values by the weights of the connections, and some output is produced.
Explain how these networks are corrected if they produce an incorrect output?
this output is compared with what it should have been,
and the difference is used to adjust the weights between the nodes
What is the more complex version of artificial neural network called?
Deep neural network
What has it already accomplished?
- it beat a professional human player at Go in 2015
- it beat the greatest living chess grandmaster in 1996
