List view
- Overdue by 3 year(s)•Due by December 4, 2022
LETS GO BABY
Overdue by 3 year(s)•Due by January 24, 2023You must hand in both a report (see below) and the source code you have developed to solve the project. For each group, only one person should hand in the project (report and code). Implementation and Code Your implementation has to be in Python. You may make use of any standard Python libraries and the numerical libraries NumPy and SciPy, as well as machine learning frameworks such as PyTorch or Tensorflow. Your code should be organised such that is easy to read, i.e. you have to use descriptive names for files, functions, variables, etc. The code may be organised in regular Python source files (.py files) or in Jupyter notebooks. If you take inspiration from or copy code developed by other people, it is important that you document this in your report. Your code must be handed in as a single file (either a zip or tar archive). A good submission would be Jupyter notebook with crisp documentation inline and all blocks being run. Report Please provide your report in pdf format. We provide a two-column template here, which you should use for your report, which must be no longer than a total of 6 pages, including figures, tables, (psoudo-)code snippets, but excluding references and appendixes. Note that the hand-in should be self-contained, although you are required to be able to explain your code for the oral exam. Your oral presentation during the exam should correspond to the structure of your write-up. The report should consist of the following sections. The goal is to be crisp, informative, and interesting, while providing the necessary facts and arguments from your experimentation. Introduction. Here you provide the context for the problem and re-state it briefly. Methods. Here you define and describe your methods, with precise mathematics where applicable. Experiments. Here you describe your data source and provide the numerical results of your method over the data. The results refer to the quantitative and qualitative evaluations that you need to carry out. Discussion. Here you reflect and discuss the results against the original question's setting and the results published in the scientific article selected. Here you also discuss peculiar strengths and short-comings of the methodology/data and argue for potential future work. Conclusions. Here you provide a couple of sentences summarising the results of the project and the indicated future improvements. References. This is a numbered reference list of the works you cite in the report, of the key methods referred. This section is not counted in the page limit. Honesty Codex By submitting your project, you confirm that you have not submitted this project elsewhere and that you did not copy (plagiarise) foreign content without proper reference. If you use someone else's work, you must cite and refer to it properly.
Overdue by 3 year(s)•Due by January 3, 2023The aim for the presentation round in class (during the last lecture/exercises slots of the semester, 08.12. from 12h ct onwards) is to inform each other about your experiences and to focus your project work on one key outcome. Thus, we organise the group presentation round as a mini-conference. Each project group is asked to: provide a 7 min. presentation + 3 min. Q/A - it is recommended to limit to 5 slides or a live notebook, give highlights of your project: central problem, central methods, key results, most important and interesting lesson learned; focus on the audience! Thus inform and motivate your fellow students of our AML course, ideally: (be prepared to) demonstrate the model/experiment/evaluation live in class for interesting discussions. Additionally, one student from the last presenting group moderates the next presenting group (thus, one student from the last group moderates the first group), which means: help with the set up and announce the presentation, help a bit with the time keeping, facilitate the Q/A round.
Overdue by 3 year(s)•Due by December 8, 2022We organise a supervision meeting halfway (on 24.11. between 12-14h) as an individual session between one of the teachers and your group. We will organise a slot of about 30min per group with a schedule beforehand (in LearnIT). The meeting is about helping you with your current open questions and challenges - code or results can be discussed but are not required.
Overdue by 3 year(s)•Due by November 24, 2022