After successful completion of the course, you will be able to understand the computational neuroscience literature, in terms of being able to implement and critically evaluate these models in terms of whether they address the underlying neuroscience problem, and write computer programs to implement a number of common computational models.
|
|
The course consists of two parts. One part is given by Kappen and consists of the following topics: stochastic neuron models and networks, deep networks and sparse coding, learning rules for classification, Boltzmann Machines and reinforcement learning. This part contains many computer excercises that are also the final examination.
The other part is taught by Tiesinga uses the tools of nonlinear dynamics, neural networks and information theory (covered in Computational Neuroscience, NWI-NM047D) to study models for cognitive processes, such as attention, decision making, memory and learning. Each year there will be a variety of topics covered in the lectures choosen from amongst modern literature on reinforcement learning, neuromorphic computing (?),neural oscillations in cognitive processes and information transfer, plasticity, reduction of large-scale models to low-dimensional dynamics, and single neuron models at various levels of complexity. The students will do small problem sets in a weekly practice hour during the course (30%) and a final project (70%) for the end grade.
|
|
|
|
Neurophysics 1 and Neurophysics 2 |
|
Written examination with practical exercises during the course, culminating in an end project. |
|
The course relies on active student participation. The students will present part of the material. The examination is comprised of written and exam and is based as well on presentations during the course, the regular and computer assignments, and on an essay that summarizes the recent developments on a particular neuroscience topic. |
|