During this seminar, students will participate in a Kaggle machine learning competition (or a similar format); read and analyze literature relating to machine learning; and present their results and findings. Possibly, this module contains an element of peer-review where students review each other’s work. Hence, students should be aware (and agree) that their source code and results will be shared among all participants. Students will work individually and/or in groups. 

More details https://unisono.uni-siegen.de/

Weekly Meetings

During the lockdown, we will have weekly meetings in Zoom https://uni-siegen.zoom.us/j/99323272467?pwd=amhYZWR1ZGs5c2VvRXBzbGFGNGd2QT09

Competition 1 (Individual)

Kaggle URL: https://www.kaggle.com/c/ml-competition-43isg3301v

Deadline: 9th of December 2020, 23:59 (German time)

Competition 2 (Group)

Kaggle URL: https://www.kaggle.com/c/ml-competition-43isg3301v-part-2 (to join the competition, you need a special URL that was shared with you via email on the 9th of December).

Deadline: 3rd of February 2021, 23:59 (German time)

Presentation (Individual; Focus on Comp. 1 and 2)

15 minutes, maximum of 12 slides. Please follow the same structure as outlined for the report + “Lessons Learned” (at least 2-3 minutes). Also, include your learnings from the first competition.

Please send the slides (as Powerpoint or PDF) to contact [at] isg.beel. org and CC to joeran.beel [AT] uni- siegen.de at least 12 hours before you will hold the presentation.

Report (Individual; Focus on Comp. 2)

The report should only focus on the second competition.

Structure

Abstract

This is a short summary of your entire report

Introduction

Introduce a) the background of your work, and clearly state b) the research problem and c) your research question/goal

Related Work

Explain and discuss critically what other researchers did to answer the research question and what their results were.

Methodology

Describe the dataset used, including the format of the data, and how the data was processed. Explain how and why you (pre-)processed the data to make it suitable for your analysis. Describe the machine learning algorithms selected and how you went about selecting appropriate values for the algorithm parameters. Present plots justifying your choices and discuss your decisions.

Results & Discussion

Present, explain, and discuss the results that you obtained. Include tables and figures where appropriate. Finally, answer clearly the research question. The answer must be based only on your own experiments and results. Discuss your results also under consideration of the related work. For instance, are your results confirming the results of related work or contradicting it?

Limitations

Discuss the limitations of your work, and what steps you would undertake next if you were to continue the project.

References

Have at least 5 references that you discuss in the “related work” section. Of these 5, at least 3 should be articles published in journals or conferences. The remaining ones may be e.g. blog posts or other websites (not Wikipedia though).

Feel free to read our academic writing guidelines.

Formalities

Page Limit

10-20 pages

Format

Please write your report in the style of an academic paper in Springer LNCS Format (Templates).

Choose between the LaTeX and MS Word template
Example of a document in Springer’s LNCS format

Deadline

10th of February 2021, 23:59

Submission

As PDF via email to contact [at] isg.beel. org and CC to joeran.beel [AT] uni- siegen.de. Please name the PFD according to your name, e.g. “JohnDoe-reportMLCompetition2020.pdf”

Marking

To pass the module, a student must pass each of the following:

  • Competition 1 (Individual)
  • Competition 2 (Group)
  • Presentation
  • Report