Completed First Draft

Requirements

Update your existing web-page.

  • Revise your existing text.
  • Remove unnecessary sections.
  • Fill in your methods section.
  • Fill in your discussion section.
    • What is the key message for your work?
    • This should include your results, analysis, and conclusions.
    • Include an ethics discussion here, if appropriate, or as a separate unrelated section afterward. If your project doesn’t touch on any ethical issues, you should write about an unrelated ethical topic of your choice. It does not need to fit the narrative.
    • You can include placeholders for text, figures, tables, etc. that you do not yet have ready.
  • Add a short reflection section.
    • What would you do differently next time?
    • How would you continue this work (e.g., what extensions would you pursue)?
  • Your introduction (or related works section) should end with a description of how your project differs from existing literature, tutorials, and examples.
  • Add a section stating what you have left to do for the final submission. This should just be a bulleted list of things to complete.
  • At least 1400 words (though I suspect some groups will have substantially more).
  • Although I’ve listed several sections, you can feel free to add any additional sections as you see fit (for example, “Future Work”).

If you want some extra insight on writing a research paper, I agree with most of Simon Peyton Jones’s information in How to write a great research paper.

You will submit a link to your web-page on gradescope.

Advice and Grading

This first draft should be a mostly complete submission. You can, of course, update it prior to the final submission, but all of your work should be complete. For the final draft, you might want to polish up charts, figures, and tables, but don’t plan to make any significant changes.

As this project is fairly open-ended, it is difficult to give a precise rubric. However, here is a list of important concepts (not all papers will address all concepts), did you

  • clearly state the thesis of your project?
  • provide clear evidence supporting your thesis?
  • provide a clear explanation as to why you could not support your thesis?
  • use proper formatting for citations?
  • add appropriate visual aids (plots, diagrams, etc.)?
  • develop a new technique from scratch?
  • create your own dataset?
  • show an exploration of your dataset (summary data, a few examples, etc.)?
  • create a significant application or demo?
  • provide a thorough analysis of the ethical implications?
  • survey several existing works and discussion their strengths and weaknesses?
  • compare several different methods (e.g., CNNs vs fully-connected vs RNNs)?
  • compare several different models (e.g., Resnets, YOLO, Inception, custom)?
  • compare several different sets of hyperparameter values?
  • compare your results with a baseline (e.g., human-level accuracy, or a non-machine learning method)?
  • compare different frameworks (e.g., PyTorch, TensorFlow, etc.)
  • create an application for your final model?
  • support your claims/analysis with appropriate data and figures?
  • create something beyond what you’d find in a simple tutorial?
  • describe the strengths and weaknesses of your data, model, and approach in general?
  • describe why your approach did not work as well as you expected?
  • discuss what you could do to improve your approach?