Student Seminar in Statistics: Statistical Learning with Sparsity
Autumn semester 2018
General information
Lecturer | Martin Mächler |
---|---|
Assistants | Yulia Kulagina, Solt Kovács |
Lectures | Mon 15-17 HG D 7.2 >> |
Course catalogue data | VVZ |
Course content
AbstractWe study selected chapters from the 2015 book "Statistical Learning with Sparsity" by Trevor Hastie, Rob Tibshirani and Martin Wainwright.
ObjectiveDuring this seminar, we will study roughly one chapter per week from the book. You will obtain a good overview of the field of sparse and high-dimensional modeling of modern statistics. Moreover, you will practice your self-studying and presentation skills.
Literature
Trevor Hastie, Robert Tibshirani, Martin Wainwright (2015):
Statistical Learning with Sparsity: The Lasso and Generalizations
Monographs on Statistics and Applied Probability 143
Chapman Hall/CRC ISBN 9781498712170
Access:
full access via ETH (library) network, if inside ETH (or VPN) or
Author's website (includes errata, updated pdf, data) .
Announcements
-
September 28th 2018:
The assignment of the topics can be found at "Course materials and schedule". -
September 7th 2018:
Welcome to the class website! The first class will take place on Monday, September 24th 2018. We are looking forward to seeing you then.
Assignment of topics: The first two topics will be assigned per email before the start of the semester. The remaining topics will be assigned in the first class. Please indicate your preferences via e-mail by 19/09/2018. Also let us know ASAP in case you decide not to take part in the seminar.
Course materials and schedule
The presentations should be roughly 50-60 minutes. The assistants will meet with you twice before your presentation, to answer questions about the material and to give feedback on your planned presentation. For more detailed guidelines for the presentations see the FAQs.
Week | Topic | Questions and Summary | Students' slides/R-code/handouts |
---|---|---|---|
Week 1 (24/09/2018) | Group 1: Chapter 2 without 2.4 and Chapter 11.1-11.3
|
|
|
Week 2 (01/10/2018) | Group 2: Chapter 11.4 and 4.1-4.4 (without some of the computational parts)
|
|
|
Week 3 (08/10/2018) | Group 3: Chapter 3 (without 3.7)
|
|
|
Week 4 (15/10/2018) | Group 4: Chapter 2.4, 5.1-5.3 and 3.7 ("computation & optimization for Lasso - part 1")
|
|
|
Week 5 (22/10/2018) | Group 5: Chapter 5.4-5.10 ("computation & optimization for Lasso - part 2")
|
|
|
Week 6 (29/10/2018) | Group 6: Chapter 4.5-4.6 (plus either remaining computational parts of Chapter 4 or change point detection via fused Lasso flavored ideas (e.g. this paper from JASA, 2010) Change point detection for high dimensional models: regression and/or covariance matrices (e.g. this paper )
|
|
|
Week 7 (05/11/2018) | Group 7: Chapter 6 (plus some example or more details on POSI via e.g.
this paper)
|
|
|
Week 8 (12/11/2018) | Group 8: "SfS approaches" to high-dimensional inference
( paper incl. R package "hdi")
|
|
|
Week 9 (19/11/2018) | Group 9: Chapter 7
|
|
|
Week 10 (26/11/2018) | Group 10: Chapter 8 (without 8.3 and some parts of 8.5)
|
|
|
Week 11 (03/12/2018) | Group 11: Chapter 9
|
|
|
Week 12 (10/12/2018) | Group 12: Chapter 10
|
|
|
Week 13 (17/12/2018) | ... |
FAQ
- What are the expectations for the seminar participants to obtain ECTS points?
- One well-prepared presentation as 2-person team, totalling 50-60 minutes.
- Two preparation meetings with the assigned assistant, ca. 2 weeks and 4-6 days before your presentation.
- Act as official discussant ("question" and "summary" group) for two of the other students' presentations. Once you will be in charge of asking questions ("question" group) and once you will give a short summary of the presentation ("summary" group). See further details later on in FAQ.
- Give a short, but constructive written feedback on strengths and weaknesses of other groups right after their presentations.
- You attend the seminar at least 10 times. In case of medical problems or other contingent extra absences you would have to present written "proof" to us.
- I am on the waiting list, when will I know if I am able to attend the course?
Currently, all 24 places are taken. If some students drop the course, the students from the waiting list will get a chance to attend. We will know the final list of students on the 2nd of October.
- When and how will the presentations be assigned?
We will assign the first 2 presentations before the start of the semester. The remaining presentations will be assigned on the 24th of September. You can indicate your preferences via e-mail by the 19th of September.
- How long should the presentation be?
The total presentation time should be 50-60 minutes. Each student should present roughly half the time. We advice you to split the presentation in two roughly equal parts. Make sure to practice so that you don't go over your time!
- Do I have to present proofs from the text book?
We encourage you to focus on getting the main ideas accross. In this respect, examples and intuition are often more useful than formal proofs. You can discuss it with the assistant.
- Do I have to present all examples from the text book?
No, you can determine which examples are best for your presentation. You can also come up with your own examples (e.g. simulation results).
- Should I use a template for my slides?
You can use any template you like. We recommend using one of the ETH presentation templates. Note that LaTeX templates are also available amongst the ETH presentation templates. We encourage you to use LaTeX.
- How should the presentation be structured?
The main purpose of the presentation is not to show what you have learned from reading the chapter, but to transmit this knowledge to the audience. So after studying the material, please take a step back and try to put yourself in the shoes of the audience: What would they already know? What would they find most interesting? What would be helpful examples? We highly encourage you to try to interact with the audience during your presentation.
You may also find it useful to read this guideline from a previous seminar. (For this seminar we do not require that you prepare a handout.) - Do I need to bring my own laptop to present my slides?
Ideally, yes. If you do not have a laptop, or you do not have a way of connecting to an hdmi or vga connector, please let the assistants know in advance.
- Will my slides be published somewhere?
Yes, all slides will be published on the course website after the presentation. Please make sure to respect copyright. In particular, if you include any images or tables not made by you in the presentation, make sure to include the source of the image/table as well.
- Do I need to prepare a handout?
This is optional. If you choose to prepare a handout, please make sure to bring 26 copies. The handout will also be posted online with your presentation.
- What is the role of the assistants?
The assistants give you guidance and feedback prior to your presentation. You will have a chance to meet with them about two weeks before the presentation for the first time and 4-6 days before your presentation for the second time. You will also get a short feedback on your performance right after your presentation. The assistants are also happy to answer other questions that are not answered in this FAQ.
- What am I supposed to do BEFORE the two meetings with the assistants?
It is expected that the student has already read the material and is familiar with the topic (i.e. at least understood the main ideas) before the first meeting. Two weeks prior to the presentation date each group attends a meeting with one of the assistants to discuss the topic and the (remaining) questions to the material, as well as an outline of the presentation.
4-6 days prior to the presentation each group meets again with the same assistant (they have to agree on a time). It is expected that presentation and handout (optional) are more or less finished. The goal of the second meeting is to finalize the presentation. - What do the "question" and "summary" groups do during/after a presentation?
The "question" group studies the chapter that is being presented in advance. They follow the presentation extra closely and take the lead in asking questions (during and/or after the presentation), for example if some terms or ideas are not clearly explained. Of course, the rest of the class should also participate acitvely and ask questions!
After the presentation and questions, the "summary" group will give a short summary of the presentation and constructive comments on the strengths and weaknesses of the presentation. - Do I get feedback after my presentation?
Other students listening to your presentation will give you short written feedback. Moreover, you will get feedback from the assistants right after your presentation.
- How do I give feedback for others?
We regard feedback as an important part of such a seminar. From a previous seminar we append the following rules (here) for giving and receiving feedback, as well as criteria (here) for feedback. These should serve as the basis for the written feedback for other groups and also for the feedback of the "summary" group.