CSE599 Research Miniproject

An individual project (self chosen) that focuses on some aspect of the class material. We want you to examine a contemporary research problem in alternative computing paradigms such as DNA computing, neuronal computing, or quantum computing. Your project may be a simulation or theoretical analysis. You must hand in a project writeup (5 to 15 pages) that clearly and precisely describes the problem and your results. You may use any software tools to help you with your project.

You will schedule a 30-minute meeting with the instructor on 6/7, 6/8, or 6/9 to describe your project and hand in your writeup.

Project ideas (partial list—more to follow)

Investigate building a reliable computer using unreliable components. Examine the early work done by von Neumann, Cowan, and others, as well as more recent analysis (and also some prototype systems). Look at and discuss the possibilities. Perhaps simulate a small model system.

Read about mesoscale quantum systems (start by reading about the quantum hall effect).  Hypothesize about the possibility of building a mesoscale quantum computer.

Does antidromic enervation (action potentials traveling back up a dendrite) make a case for backpropagation learning in nervous tissue? Investigate the sensitivity of the conventional backpropagation-of-errors algorithm to roundoff errors or noise. Can backprop work at low resolution or with poor node selectivity?

Investigate the rate-code versus temporal-code arguments from biology. Can neurons transmit more information using a rate code or a temporal code? How might you encode information in unclocked spike trains? Perhaps simulate a small spike-based pattern-recognition system.

Simulate the execution of a stored program using self-assembly of DNA (i.e. simulate solving a toy problem using DNA). Include binding errors and the forward and backward reaction rates in your simulation.

Investigate contemporary research in adiabitic computing. Analyze the possibilities, or simulate the benefits when compared with conventional computing.

Comments to: cse599-webmaster@cs.washington.edu