DSS5104 Assignment Topics
DSS5104
Here is a list of possible topics for your assignment.
- Molecular Properties Prediction
- Deep Learning for Time-Series Forecasting
- Deep Learning for Tabular Data
- Transfer Learning for Computer Vision
- Deep Learning for Recommender Systems
- Fraud and Anomaly Detection with Deep Learning
- Text Classification: From Classical NLP to Transformers
- Deep Learning for Image Segmentation
- Open Topic
Assignment group: You are reminded that assignments are to be done in groups of at most 4 students: each student in a group will be receiving the same mark. In all your reports, do include the list of all students, as well as their student numbers.
Submission instructions:
- Submit your report as a PDF file on CANVAS. Only PDF files are accepted; no other formats will be graded.
- Your code must be in a GitHub repository. Include the link to your repository in your report. Do not submit code on CANVAS. All results stated in your report must be reproducible from the repository.
Late submissions will incur a penalty of \(10\%\) per day late: if you original mark is, say, \(81/100\), and you are \(6\) days late, your mark will be \(81 \times 0.9^6 = 43\).