EEBM E9070: Massively Parallel Neural Computation
| Lecturer: | Professor Aurel A. Lazar | |
| Office hours: | By appointment, Room 819 Schapiro | |
| E-mail address: | aurel "at" ee.columbia.edu | |
| Class Web Site: | Offered by CourseWorks | |
| Day and Time: | Tuesdays, 6:50 PM - 8:40 PM | |
| Class Location: | EE Conference Room (1300 Mudd) | |
| Credits for course: | 3 points | |
| Prerequisites | BMEB W4020 or the instructor's approval | |
| Description: | Drosophila connectomics. Detailed description of the fruit fly's olfactory and vision systems. Parallel Processing on GPUs. | |
| Recommended texts: | --- David B. Kirk and Wen-mei W. Hwu,
Programming Massively Parallel Processors: A Hands-on Approach,
Morgan Kaufmann, 2010. NVIDIA CUDA, NVIDIA CUDA C Programming Guide, Version 4.0, 2011. |
|
| Homework(s): | Reading research papers and book chapters | |
| Paper(s): | --- | |
| Project(s) | Implementation of processing units of the fruit fly's brain in Python/PyCUDA. (Strongy motivated students with prior background in Neuroscience and Parallel Computing can request working on an individual project - instructor's approval required.) | |
| Midterm exam: | Project I: Massive connectivity and feedback in neural circuits. | |
| Final Exam: | Project II: Architecture of a sensory system of your choice (instructor's approval required). | |
| Grading | Classroom participation and projects | |
| Hardware requirements: | Laptop for demos. | |
| Software requirements: | Remote access to GPUs running Python and PyCUDA will be provided. | |