Projects in Massively Parallel Neural Computation
Students will utilize the power of GPUs to implement massively parallel simulations of neural circuits and architectures of the fruit fly brain. You will be closely involved in existing or new projects in fruit fly vision, olfaction, audition or locomotion. All projects provide students with a hands-on opportunity to learn about and use the latest Fermi or Kepler GPUs from NVIDIA.
Projects pertain to
- spike processing of visual stimuli; students will be applying knowledge of image processing and computer vision/graphics to neural processing.
- neural coding of spatiotemporal signals (auditory, olfactory and visual); students will be applying knowledge of information/communication/coding theory to neural encoding and decoding.
These projects require introductory background (undergraduate/graduate level) in both computational neuroscience and in one of the concentrations (subfields) of Electrical/Computer Engineering or Computer Science.
Students with extensive scientific computing experience are strongly encouraged to apply.
Ideally, qualified students meet the following (guiding) coursework requirements. Students with backgrounds in other relevant fields/departments may be accepted after assesment by Prof. Lazar.
Qualified students must meet the following requirements.
- BMEB W4020 - Computational Neuroscience or equivalent.
In addition, students must have taken at least two courses from anyone of the following four groups of courses
- ELEN E4830 - Digital Image Processing or equivalent.
- COMS W4731 - Computer Vision or equivalent.
- COMS W4160 - Computer Graphics or equivalent.
- ELEN E6712 - Communication Theory or equivalent.
- ELEN E6717 - Information Theory or equivalent.
- ELEN E6718 - Algebraic Coding Theory or equivalent.
- ELEN E4312 - Analog Electronic Circuits or equivalent.
- ELEN E4321 - Digital VLSI Circuits or equivalent.
- CSEE W4824 - Computer Architecture or equivalent.
- COMS W4771 - Machine Learning or equivalent.
- ELEN E6010 - Systems Biology or equivalent.
- ELEN E6860 - Advanced Digital Signal Processing or equivalent.
- Knowledge of Matlab is required.
- Python, and/or C/C++ is strongly recommended.
- Knowledge of NVIDIA's CUDA and/or OpenCL parallel programming environments is a plus.
Qualified students with programming experience who have not worked with Matlab may be accepted after assessment by Prof. Lazar.