CO673/CS794 - Assignments

5 assignments will be posted biweekly (see below for rough dates; small variations may apply).

Grade breakdown: 5 x 20% = 100%.

Please submit your completed homework electronically through LEARN. Submit early and often!

We do not accept late submissions (except when you have a legitimate reason).

Grade appeal must be made within 2 weeks of receiving the grade.

All due dates are 11:59pm AOE (anywhere on earth) or equivalently 08:00am EST (+1 day)

• assignment 5 [pdf, tex]: out Nov 20; due 11:59pm, Dec 04

  • default dataset website; 30k instances, 23 features, 2 labels (0 or 1)

  • pickle file and jupyter notebook to unpack and load data

  • original dataset and jupyter notebook to load data into numpy array

• assignment 4 [pdf, tex]: out Nov 06; due 11:59pm, Nov 20

Alternatively, you may choose to substitute the last assignment with a course project, such as the following:

• a survey of a class of optimization algorithms for data science and machine learning,

• an empirical comparison of several related algorithms on an interesting dataset,

• an application of the discussed algorithms to a new field/application,

• a novel algorithm to address a need in machine learning / data science,

• a (new or improved) theoretical analysis of an optimization algorithm (new or old),

• translate a classical paper of the instructor's choosing and discuss the relevant literature (German, Italian, or Russian).

Instructor approval is required for the project. It would be wise to complete the first 4 assignments and focus on your project afterwards.

Please use the following tex files to write your final report:

• tex

• style

• sample bib

• sample pdf