DS-GA 1018: Probabilistic Time Series Analysis

Project Expectations

Course: DS-GA 1018: Probabilistic Time Series Analysis
Due: Wednesday, November 26th

Project presentation: 50%

Presented in class in the last session. The project does not need to be finished by this point, you can just present your progress.

Format: A Google slide deck for all presentations (to be sent by the course instructor)

Time limit: Max 5 minutes presentation per group, then 2-3 minutes Q/A from the course instructor

Content: Should cover motivation/summary and how it’s situated within related literature, key contributions you want to make, your progress so far (models, experiments, results, etc), what you plan to do next.

Grading: You will be graded on the three criteria outlined below along with guiding questions explaining what graders will be looking for:


Project report: 50%

Compose this as you would a research paper. While the exact structure is up to you, a typical paper will have an abstract, introduction, related works, methods, experiments, results, discussion, and conclusion. Aim for about 4 pages + references.

Templates: As you like, 12pt font, single-spaced. The instructor will update this with example templates.

Some tips on writing ML research papers:
- Grigorios G. Chrysos - How to write a research paper in machine learning
- ICML 2002 - How to write a great research paper
- Zack Lipton - Heuristics for Scientific Writing (A Machine Learning Perspective)
- John Wentworth - Highly opinionated advice on how to write ML papers

Grading - you will be graded on the three criteria outlined below along with guiding questions explaining what graders will be looking for: