11-785 Introduction to Deep Learning
Fall 2019

Poster Guidelines

What Dimension?

The size of the poster should be 30x40in or 40x30in as in template.

Where to print?

Where to submit?

One pdf poster document per team should be submitted on Canvas on December 3rd by no later than 11:59pm.

Presentation Date and Location

Poster Presentatons will be held on Thursday December 5th in GHC4300 Commons and Monday December 9th in GHC7107 Atrium from 2:00pm to 5:00pm

What to Include in Your Poster?

Be creative, after all it is a presentation! You should be able to illustrate what your project can do.
Overall, your poster should be aligned with these recommendations:

key point Description
Overall Clarity of poster Define clear sections in your poster.
It should present the key points of your work, up to where you got with it.
Summarize ideas and consider using figures and tables to make it more illustrative as that will give a better sense of work done.
Introduction What are you trying to solve? Why is it important in the field?
Related work Previous work related to your topic that served as a baseline or motivation.
This would serve as your backup support in what you have developed.
Methods Which techniques did you apply?
You are strongly encouraged to add an illustration of your Machine Learning pipeline.
(suggested: figures)
Dataset Description of the dataset used, number of files, train and test partitions, labels, etc.
(suggested: tables)
Results Your experimental results.
What metrics did you use for evaluation?
How do your results compare to prior work or your baseline?
Conclusions summary analysis of your model and results.
Describe current results in each of the experiments aligned with your expectations.
Highlight a few limitations of your approach
(e.g., strong assumptions you had to make, constraints, when your method didn’t work in practice, etc.).
References Add the most relevant citations only

Presentation Grading Rubric

Your project will be scored according to the following rubric:

Problem Proposed Methods Completeness Results Publishability
Whether the problem description is clear?
Is the motivation strong?
Is the problem nontrivial?
Clear enough? Technically sound? Originality? Is your work complete?
Have you obtained results consistent with successful completion of the work?
Are the results reasonable?
how do you set up their experiments?
Comparisons with others? Can you explain the results?
Have you looked into the results and gained some insights on DL models/algorithms?
Is your work potentially publishable?
10% 10% 10% 10% 10%