...
2. Footnote: We make an exception for sources that can be taken for granted in the instructional setting, namely, the course materials. To minimize documentation effort, we also do not expect you to credit the course staff for ideas you get from them, although it’s nice to do so anyway.
Grading Guidelines: (still under construction)
You are allowed to use methods from outside of what is covered in class. But if you do, provide a clear comparison with ``reasonable methods’’ covered in class.
Visualization (10%)
Inclusion of plots/diagrams (5%)
Explanation of how visuals helped develop the model (5%)
Algorithms (30%)
Correct use of algorithms (15%)
Used principled approach to extract information from similarity graph and provided clear explanation and reasoning (5%)
Extracted and used common information from both the features and the similarity graph in a principled fashion and provided clear explanation and reasoning(5%)
Used clustering algorithms to cluster datapoints into classes, clearly explained and analyzed the method (5%)
Explanation of how algorithms helped to develop model (15%)
Showed evident understanding of each algorithm used
Model (40%)
Use of data (30%)
Individual testing (10%)
Tested performance on just features, just graph
Combining data (10%)
Combined data from features and graph to develop model
Partial supervision (10%)
Used seeds to classify points into classes
Parameters (10%)
Evident testing of different parameters (5%)
Reasons for choosing certain parameters (5%)
Failed Attempts (20%)
Explanation (10%)
Explained how they developed their failed models and why they think those models failed
Improvement (10%)
Explained how failed attempts led them to develop their final model
Bonus (at the discretion of the graders):
Tried new or more methods not necessarily covered in class
Developed new algorithm or methods, tweaked existing methods to fit the problem better
…