This page describes the guidelines to help you write up the final report of your project. Please note that the report should provide all the required information to evaluate your project, and at the same time, it should be concise and data-driven. I recommend you to follow these guidelines. Nonetheless, there is some flexibility, and you are welcome to add new sections as long as you remain within the general framework.
Your final report should not exceed eight single-spaced pages, using 11 pt font.
NO HANDWRITTEN report will be accepted.
The report must be written in English.
Please describe the problem you are working on and provide a brief background of it. It is recommended that you briefly discuss the traditional ways of solving it, as well as some related works if you are aware of them. Present the motivation to work on the selected problem; i.e., why is it important? How have you addressed it? Finally, summarizes the main results.
Formally define the problem you are addressing. It means, formally specifies the inputs and outputs. Elaborate on why this is an exciting and essential machine learning problem.
Describe in reasonable detail the learning model you have used to address the problem. Give the pseudocode of the learning algorithm, including its general outline and the main equations. Trace through a concrete example of how the algorithm works. The example should be complex enough to illustrate all the essential aspects of the problem, but simple enough to be easily understood. If possible, an intuitively and meaningful example is better that one with meaningless symbols.
What criteria have you used to evaluate your learning model? What hypotheses do your experiment test?
What dataset have you used to train your learning model? What are the dependent and the independent variables? If you used training, test, and validation sets, how did you split the data between these sets?
Describe the quantitative results of your experiments. A graphical representation is usually better than tables. What were the fundamental differences revealed in the data? Are they statistically significant?
Are your hypotheses supported? What conclusions do the results support the strengths and weakness of your learning model compared to other learning approaches? How can the results be explained in terms of the underlying properties of the learning algorithm and/or the data?
What are the significant shortcomings of your current learning model? For each one, propose additions or enhancements that would help overcome it?
Summarize significant results and conclusions presented in the report. What were the most critical points illustrated by your work?
Include a comprehensive list of the references cited in the report.