Statistical learning theory 2022 — различия между версиями
Bbauwens (обсуждение | вклад) |
Bauwens (обсуждение | вклад) |
||
(не показаны 72 промежуточные версии 2 участников) | |||
Строка 3: | Строка 3: | ||
− | Lectures: Friday 16h20 -- 17h40, [https://www.hse.ru/en/org/persons/160550073 Bruno Bauwens], [https://www.hse.ru/staff/mkaledin Maxim Kaledin] | + | Lectures: Friday 16h20 -- 17h40, [https://www.hse.ru/en/org/persons/160550073 Bruno Bauwens], [https://www.hse.ru/staff/mkaledin Maxim Kaledin], room M202 and on [https://us02web.zoom.us/j/82300259484?pwd=NWxXekxBeE5yMm9UTmwvLzNNNGlnUT09 zoom] |
− | Seminars: | + | Seminars: Saturday 14h40 -- 16h00, [https://www.hse.ru/org/persons/225526439 Artur Goldman], room M202 and on zoom (the link will be in telegram) |
− | + | To discuss the materials, join the [https://t.me/+G0VKOE2-nnkwNDE0 telegram group] The course is similar to [http://wiki.cs.hse.ru/Statistical_learning_theory_2021 last year]. | |
− | |||
− | == | + | == Problems exam == |
− | + | December 21, 13h-16h, computer room G403 ([https://us02web.zoom.us/j/82300259484?pwd=NWxXekxBeE5yMm9UTmwvLzNNNGlnUT09 zoomlink] for students abroad)<br> | |
+ | -- You may use handwritten notes, lecture materials from this wiki (either printed or through your PC), Mohri's book <br> | ||
+ | -- You may not search on the internet or interact with other humans (e.g. by phone, forums, etc) | ||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
== Course materials == | == Course materials == | ||
Строка 64: | Строка 57: | ||
|| [https://www.dropbox.com/s/yurv5s42w3kw5vv/04sol.pdf?dl=0 solutions 4] | || [https://www.dropbox.com/s/yurv5s42w3kw5vv/04sol.pdf?dl=0 solutions 4] | ||
|- | |- | ||
− | | [https:// | + | | [https://www.youtube.com/watch?v=8J5B9CCy-ws 30 Sept] |
|| Growth functions, VC-dimension and the characterization of sample comlexity with VC-dimensions | || Growth functions, VC-dimension and the characterization of sample comlexity with VC-dimensions | ||
|| [https://www.dropbox.com/s/rpnh6288rdb3j8m/05slides.pdf?dl=0 sl05] | || [https://www.dropbox.com/s/rpnh6288rdb3j8m/05slides.pdf?dl=0 sl05] | ||
− | || [https://www.dropbox.com/s/ | + | || [https://www.dropbox.com/s/eurz2vkvt1wa5zm/07book_growthFunctions.pdf?dl=0 ch07] [https://www.dropbox.com/s/m7xe7k39qzmzapv/08book_VCdimension.pdf?dl=0 ch08] |
− | || list 5 | + | || [https://www.dropbox.com/s/u1vpi28gwf0zig4/05sem.pdf?dl=0 list 5] |
− | || | + | || [https://www.dropbox.com/s/3sq7yzv7v4l9tbb/05sol.pdf?dl=0 solutions 5] |
|- | |- | ||
| [https://drive.google.com/file/d/17zynIg_CZ6cCNBig5QXmBx7VFS8peyuU/view?usp=sharing 07 Oct] | | [https://drive.google.com/file/d/17zynIg_CZ6cCNBig5QXmBx7VFS8peyuU/view?usp=sharing 07 Oct] | ||
|| Risk decomposition and the fundamental theorem of statistical learning theory | || Risk decomposition and the fundamental theorem of statistical learning theory | ||
− | || [https://www.dropbox.com/s/ | + | || [https://www.dropbox.com/s/0p8r5wgjy1hlku2/06slides.pdf?dl=0 sl06] |
− | || [https://www.dropbox.com/s/ | + | || [https://www.dropbox.com/s/8c87619ewkyod4f/09book_riskBounds.pdf?dl=0 ch09] |
− | || list 6 | + | || [https://www.dropbox.com/s/eyfczsuwz60moj7/06sem.pdf?dl=0 list 6] |
− | || | + | || [https://www.dropbox.com/s/1te4fzlwj72v6ph/06sol.pdf?dl=0 solutions 6] |
|- | |- | ||
− | | 14 Oct | + | | [https://www.youtube.com/watch?v=yMsUH1brAs8 14 Oct] |
− | || Bounded differences inequality, Rademacher complexity, symmetrization, contraction lemma | + | || Bounded differences inequality, Rademacher complexity, symmetrization, contraction lemma, [https://www.dropbox.com/s/8uravgo5shas55g/07quiz.pdf?dl=0 quiz] |
|| [https://www.dropbox.com/s/kfithyq0dgcq6h8/07slides.pdf?dl=0 sl07] | || [https://www.dropbox.com/s/kfithyq0dgcq6h8/07slides.pdf?dl=0 sl07] | ||
− | || [https://www.dropbox.com/s/ | + | || [https://www.dropbox.com/s/fg4seoqjbeb7a5g/10book_measureConcentration.pdf?dl=0 ch10] [https://www.dropbox.com/s/hfrvhebbsskbk6g/11book_RademacherComplexity.pdf?dl=0 ch11] |
− | || list 7 | + | || [https://www.dropbox.com/s/qofutar8qy5y53i/07sem.pdf?dl=0 list 7] <span style="color:red">update 15.10</span> |
− | || | + | || [https://www.dropbox.com/s/w1ud2okt120vm5j/07sol.pdf?dl=0 solutions 7] |
|- | |- | ||
| | | | ||
Строка 90: | Строка 83: | ||
| [https://drive.google.com/file/d/1L-BeDxhoHcoDrdlVTlfoMFwnWXKV46cr/view?usp=sharing 21 Oct] | | [https://drive.google.com/file/d/1L-BeDxhoHcoDrdlVTlfoMFwnWXKV46cr/view?usp=sharing 21 Oct] | ||
|| Simple regression, support vector machines, margin risk bounds, and neural nets | || Simple regression, support vector machines, margin risk bounds, and neural nets | ||
− | || [https://www.dropbox.com/s/ | + | || [https://www.dropbox.com/s/oo1qny9busp3axn/08slides.pdf?dl=0 sl08] |
− | || [https://www.dropbox.com/s/ | + | || [https://www.dropbox.com/s/573a2vtjfx8qqo8/12book_regression.pdf?dl=0 ch12] [https://www.dropbox.com/s/jaym44fmif2uw05/13book_SVM.pdf?dl=0 ch13] |
− | || list 8 | + | || [https://www.dropbox.com/s/bzo6msrxcfa8tpp/08sem.pdf?dl=0 list 8] |
− | || | + | || [https://www.dropbox.com/s/pe7yctcr93yaw95/08sol.pdf?dl=0 solutions 8] |
|- | |- | ||
| [https://youtu.be/9FhFxLHR4eE 04 Nov] | | [https://youtu.be/9FhFxLHR4eE 04 Nov] | ||
|| Kernels: RKHS, representer theorem, risk bounds | || Kernels: RKHS, representer theorem, risk bounds | ||
− | || [https://www.dropbox.com/s/ | + | || [https://www.dropbox.com/s/jst60ww8ev4ypie/09slides.pdf?dl=0 sl09] |
− | || [https://www.dropbox.com/s/ | + | || [https://www.dropbox.com/s/ply602zthd7r3jv/14book_kernels.pdf?dl=0 ch14] |
− | || list 9 | + | || [https://www.dropbox.com/s/54xdufimavhd646/09sem.pdf?dl=0 list 9] |
− | || | + | || [https://www.dropbox.com/s/i3rx26ya6kvm5p2/09sol.pdf?dl=0 solutions 9] |
|- | |- | ||
− | | [https://youtu.be/ | + | | [https://youtu.be/1oUXZy6Sqlk 11 Nov] |
|| AdaBoost and the margin hypothesis | || AdaBoost and the margin hypothesis | ||
|| [https://www.dropbox.com/s/umum3kd9439dt42/10slides.pdf?dl=0 sl10] | || [https://www.dropbox.com/s/umum3kd9439dt42/10slides.pdf?dl=0 sl10] | ||
− | || | + | || [https://www.dropbox.com/s/e7m1cs7e8ulibsf/15book_AdaBoost.pdf?dl=0 ch15] |
− | || list 10 | + | || [https://www.dropbox.com/s/nu4a55qbfqlp3bl/10sem.pdf?dl=0 list 10] |
− | || | + | || [https://www.dropbox.com/s/t64mjapdzcm1313/10sol.pdf?dl=0 solutions 10] |
|- | |- | ||
− | | 18 Nov | + | | [https://youtu.be/GL574ljefJ8 18 Nov] |
|| Implicit regularization of stochastic gradient descent in neural nets | || Implicit regularization of stochastic gradient descent in neural nets | ||
|| | || | ||
− | || | + | || [https://www.dropbox.com/s/b4xac5uki7l1ysq/16book_implicitRegularization.pdf?dl=0 ch16] |
− | || | + | || no seminar |
|| | || | ||
|- | |- | ||
Строка 119: | Строка 112: | ||
|| ''Part 4. Other topics'' | || ''Part 4. Other topics'' | ||
|- | |- | ||
− | | 25 Nov | + | | [https://youtu.be/kOXi_m9dBzE 25 Nov] |
− | || Regression I: | + | || Regression I: fixed design with sub-Gaussian noise |
− | + | ||
− | + | ||
− | + | ||
|| | || | ||
+ | || [https://disk.yandex.ru/i/tI7NiGsvQP0Jww notes12] | ||
+ | || [https://disk.yandex.ru/d/9fFxVlMw4kPfEQ list 12] | ||
+ | || [https://disk.yandex.ru/i/5vBE2VC7zNC3Rg solutions 12] | ||
|- | |- | ||
− | | 02 Dec | + | | [https://youtu.be/GEYT_IxXEX0 02 Dec] |
− | || | + | || Multiarmed bandids I |
|| | || | ||
− | || | + | || [https://disk.yandex.ru/i/lvqXofEbaFkfAA notes13] |
− | || list 13 | + | || [https://disk.yandex.ru/i/ZXXJbBiJUPNiOw list 13] |
|| | || | ||
|- | |- | ||
− | | 09 Dec | + | | [https://youtu.be/Uybf6mCp2Es 09 Dec] |
− | || Multiarmed bandids | + | || Multiarmed bandids II (optional) |
− | + | ||
|| | || | ||
− | || | + | || [https://disk.yandex.ru/i/Nqy9-wmZ-g5o8g notes14] |
+ | || [https://disk.yandex.ru/d/0Qupo2CNSjS_pQ notebook], [https://disk.yandex.ru/d/suY6d58SFf09Bg notebook(solved)] | ||
|| | || | ||
|- | |- | ||
Строка 148: | Строка 141: | ||
Foundations of machine learning 2nd ed, Mehryar Mohri, Afshin Rostamizadeh, and Ameet Talwalker, 2018. This book can be downloaded from [https://libgen.is Library Genesis] (the link changes sometimes and sometimes vpn is needed). | Foundations of machine learning 2nd ed, Mehryar Mohri, Afshin Rostamizadeh, and Ameet Talwalker, 2018. This book can be downloaded from [https://libgen.is Library Genesis] (the link changes sometimes and sometimes vpn is needed). | ||
+ | <!-- A gentle introduction to the materials of the first 3 lectures and an overview of probability theory, can be found in chapters 1-6 and 11-12 of the following book: | ||
+ | Sanjeev Kulkarni and Gilbert Harman: An Elementary Introduction to Statistical Learning Theory, 2012.--> | ||
+ | == Grading formula == | ||
− | = | + | Final grade = 0.35 * [score of homeworks] + 0.35 * [score of colloquium] + 0.3 * [score on the exam] + bonus from quizzes. |
+ | |||
+ | All homework questions have the same weight. Each solved extra homework task increases the score of the final exam by 1 point. | ||
− | + | There is no rounding except on the final grade. Grades fractional grades above 5/10 are rounded up, those below 5/10 are rounded down. | |
− | + | Autogrades: if you only need 4/10 to pass with maximal final score, it will be given automatically. This may happen because of extra questions and bonuses from quizzes. | |
− | + | ||
− | + | ||
+ | For students who want to pass with 4/10 with minimal effort: each year on the exam, I ask to calculate the VC-dimension or Rademacher complexity of some class. It should be easy to have 4/10 for the final exam. If you understand all lecture notes, you pass the colloquium with maximal score. Together this is enough. If only a few students fail and the grades are at least 3.8/10 then failed students may resubmit a few homework tasks to pull up the grade. (This happened in the last 3 years.) | ||
− | <!-- | + | |
− | + | == Colloquium == | |
+ | |||
+ | [https://www.dropbox.com/s/7djya6nc8ietd32/colloqQuest.pdf?dl=0 Rules and questions.] Update 12/12 added question 24 and corrected typos. | ||
+ | |||
+ | <!-- [https://docs.google.com/spreadsheets/d/13ox_EN6YJBEC93A6YgbbzawXE2RfynyQTUf90SXE4GQ/edit?usp=sharing Choose the day: 16 or 17 Dec.] --> | ||
+ | |||
+ | == Homeworks == | ||
+ | |||
+ | Email to brbauwens-at-gmail.com. Start the subject line with SLT-HW. | ||
+ | |||
+ | Deadline before the start of the lecture, every other lecture. | ||
+ | |||
+ | Sat. 17 Sept 18h10: problems 1.7, 1.8, 2.9, and 2.11 <br> | ||
+ | Sat. 01 Oct 18h10: see lists 3 and 4, and 2.10 <br> | ||
+ | Fri. 14 Oct 16h20: see problem lists 5 and 6 <br> | ||
+ | Sat. 05 Nov 20h00: see problem lists 7 and 8 <br> | ||
+ | Sat. 19 Nov 20h00: see problem lists 9 and 10 <br> | ||
+ | Sun. 04 Dec 23h59: see problem list 12 send it to maxkaledin@gmail.com with subject line SLT-HW-Reg <YourName>_<YourSurname> | ||
+ | <br> | ||
Строка 171: | Строка 186: | ||
| Bruno Bauwens || 15-20h || || || || 18-20h || | | Bruno Bauwens || 15-20h || || || || 18-20h || | ||
|- | |- | ||
− | | Maxim Kaledin || | + | | Maxim Kaledin || Write || in || Telegram || time is || flexible || |
|- | |- | ||
|} | |} |
Текущая версия на 19:22, 4 сентября 2023
Содержание
General Information
Lectures: Friday 16h20 -- 17h40, Bruno Bauwens, Maxim Kaledin, room M202 and on zoom
Seminars: Saturday 14h40 -- 16h00, Artur Goldman, room M202 and on zoom (the link will be in telegram)
To discuss the materials, join the telegram group The course is similar to last year.
Problems exam
December 21, 13h-16h, computer room G403 (zoomlink for students abroad)
-- You may use handwritten notes, lecture materials from this wiki (either printed or through your PC), Mohri's book
-- You may not search on the internet or interact with other humans (e.g. by phone, forums, etc)
Course materials
Video | Summary | Slides | Lecture notes | Problem list | Solutions |
---|---|---|---|---|---|
Part 1. Online learning | |||||
02 Sept | Philosophy. The online mistake bound model. The halving and weighted majority algorithms movies | sl01 | ch00 ch01 | list 1 update 05.09 | solutions 1 |
09 Sept | The perceptron algorithm. The standard optimal algorithm. | sl02 | ch02 ch03 | list 2 update 25.09 | solutions 2 |
16 Sept | Kernels and the kernel perceptron algorithm. Prediction with expert advice. Recap probability theory. | sl03 | ch04 ch05 | list 3 | solutions 3 |
Part 2. Distribution independent risk bounds | |||||
23 Sept | Sample complexity in the realizable setting, simple examples and bounds using VC-dimension | sl04 | ch06 | list 4 | solutions 4 |
30 Sept | Growth functions, VC-dimension and the characterization of sample comlexity with VC-dimensions | sl05 | ch07 ch08 | list 5 | solutions 5 |
07 Oct | Risk decomposition and the fundamental theorem of statistical learning theory | sl06 | ch09 | list 6 | solutions 6 |
14 Oct | Bounded differences inequality, Rademacher complexity, symmetrization, contraction lemma, quiz | sl07 | ch10 ch11 | list 7 update 15.10 | solutions 7 |
Part 3. Margin risk bounds with applications | |||||
21 Oct | Simple regression, support vector machines, margin risk bounds, and neural nets | sl08 | ch12 ch13 | list 8 | solutions 8 |
04 Nov | Kernels: RKHS, representer theorem, risk bounds | sl09 | ch14 | list 9 | solutions 9 |
11 Nov | AdaBoost and the margin hypothesis | sl10 | ch15 | list 10 | solutions 10 |
18 Nov | Implicit regularization of stochastic gradient descent in neural nets | ch16 | no seminar | ||
Part 4. Other topics | |||||
25 Nov | Regression I: fixed design with sub-Gaussian noise | notes12 | list 12 | solutions 12 | |
02 Dec | Multiarmed bandids I | notes13 | list 13 | ||
09 Dec | Multiarmed bandids II (optional) | notes14 | notebook, notebook(solved) | ||
16 Dec | Colloquium |
The lectures in October and November are based on the book: Foundations of machine learning 2nd ed, Mehryar Mohri, Afshin Rostamizadeh, and Ameet Talwalker, 2018. This book can be downloaded from Library Genesis (the link changes sometimes and sometimes vpn is needed).
Grading formula
Final grade = 0.35 * [score of homeworks] + 0.35 * [score of colloquium] + 0.3 * [score on the exam] + bonus from quizzes.
All homework questions have the same weight. Each solved extra homework task increases the score of the final exam by 1 point.
There is no rounding except on the final grade. Grades fractional grades above 5/10 are rounded up, those below 5/10 are rounded down.
Autogrades: if you only need 4/10 to pass with maximal final score, it will be given automatically. This may happen because of extra questions and bonuses from quizzes.
For students who want to pass with 4/10 with minimal effort: each year on the exam, I ask to calculate the VC-dimension or Rademacher complexity of some class. It should be easy to have 4/10 for the final exam. If you understand all lecture notes, you pass the colloquium with maximal score. Together this is enough. If only a few students fail and the grades are at least 3.8/10 then failed students may resubmit a few homework tasks to pull up the grade. (This happened in the last 3 years.)
Colloquium
Rules and questions. Update 12/12 added question 24 and corrected typos.
Homeworks
Email to brbauwens-at-gmail.com. Start the subject line with SLT-HW.
Deadline before the start of the lecture, every other lecture.
Sat. 17 Sept 18h10: problems 1.7, 1.8, 2.9, and 2.11
Sat. 01 Oct 18h10: see lists 3 and 4, and 2.10
Fri. 14 Oct 16h20: see problem lists 5 and 6
Sat. 05 Nov 20h00: see problem lists 7 and 8
Sat. 19 Nov 20h00: see problem lists 9 and 10
Sun. 04 Dec 23h59: see problem list 12 send it to maxkaledin@gmail.com with subject line SLT-HW-Reg <YourName>_<YourSurname>
Office hours
Person | Monday | Tuesday | Wednesday | Thursday | Friday | |
---|---|---|---|---|---|---|
Bruno Bauwens | 15-20h | 18-20h | ||||
Maxim Kaledin | Write | in | Telegram | time is | flexible |
It is always good to send an email in advance. Questions and feedback are welcome.