Marking criteria for COMP9444 Group Project
Total marks for the project work: 30 marks.
1. Project Notebook(s): 15 Marks
2. Summary Report (max 4 pages excluding references): 5 marks
3. Project Presentation: 10 Marks
Breakdown of marks for each component:
1. Marking criteria for Project notebook(s) [Total 15 marks]
Introduction, Motivation and/or Problem Statement (1 mark): Clearly define the problem statement or purpose of the project.
Data Sources or RL Tasks (1 mark): Data sources or reinforcement learning tasks are clearly documented and described.
Exploratory Analysis of Data or RL Tasks (3 marks): Provide details about the properties, number of classes, pre-processing, challenging aspects, etc. of the data (or the RL task).
Models and/or Methods (3 marks): Model(s) and/or Method(s) are judiciously chosen and appropriately applied. If building on previous work, identify the source and clearly delineate which parts are your own work.
Results (3 marks): Results are clearly shown, discussed, evaluated using appropriate metrics. Good use of graphs or other visualizations, where possible. Comparison with previous methods/State-of-the- art, where appropriate.
Discussion (2 marks): Discuss the results and analysis, provide some insight about system performance, including strengths, weaknesses, limitations and possible future work.
Writing (2 marks): Notebook(s) are presented in a readable format, appropriate section/subsection headings are provided in markdown format, codebase is easy to follow.
2. Marking criteria for Project Summary Report [Total 5 marks]
Introduction (0.5 mark): Provide a high-level description of the project.
Literature Review (0.5 mark): Review existing methods/techniques relevant to the project.
Models and/or Methods (0.5 mark): Justify and explain the selection of models and/or methods appropriate to the task.
Experimental Setup (0. 5 mark): Provide clear details about model parameters, evaluation metrics to be used, data split into training, validation, and testing, etc.
Results (0.5 mark): Discuss the main findings and results. How well does the system perform? Compare to other method(s)/SOTA, if possible.
Conclusions (0.5 mark): What are the key strengths and weaknesses of the proposed solution? What are the key limitation(s)? Recommendations for future work.
Details in the report (1 mark): Report provide sufficient details to understand the project clearly - motivation, dataset, rationale for the selection of models, results, and conclusions.
Overall quality of the report (1 mark): Summary Report is well-presented and formatted, within specified length, without grammatical mistakes or typos.
3. Marking criteria for Project Presentation [Total 10 marks]
Individual-based
Contribution to the presentation (2.5 marks)
Clearly spoken and understandable (2.5 marks)
Group-based
Clear problem statement (1 mark)
Clear presentation of data and data exploration or RL task (1 mark)
Clear presentation of models and/or methods (1 mark)
Clear presentation of results, result analysis or error analysis/qualitative examples (1 mark)
Quality of slides or visual presentation material (1 mark)