<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="ru">
	<id>https://wiki.cs.hse.ru/api.php?action=feedcontributions&amp;feedformat=atom&amp;user=Storandrew</id>
	<title>Wiki - Факультет компьютерных наук - Вклад [ru]</title>
	<link rel="self" type="application/atom+xml" href="https://wiki.cs.hse.ru/api.php?action=feedcontributions&amp;feedformat=atom&amp;user=Storandrew"/>
	<link rel="alternate" type="text/html" href="https://wiki.cs.hse.ru/%D0%A1%D0%BB%D1%83%D0%B6%D0%B5%D0%B1%D0%BD%D0%B0%D1%8F:%D0%92%D0%BA%D0%BB%D0%B0%D0%B4/Storandrew"/>
	<updated>2026-09-21T15:32:59Z</updated>
	<subtitle>Вклад</subtitle>
	<generator>MediaWiki 1.43.9</generator>
	<entry>
		<id>https://wiki.cs.hse.ru/index.php?title=Theory_of_Computing_2018_2019&amp;diff=29471</id>
		<title>Theory of Computing 2018 2019</title>
		<link rel="alternate" type="text/html" href="https://wiki.cs.hse.ru/index.php?title=Theory_of_Computing_2018_2019&amp;diff=29471"/>
		<updated>2018-09-26T12:23:10Z</updated>

		<summary type="html">&lt;p&gt;Storandrew: &lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== General Information ==&lt;br /&gt;
&lt;br /&gt;
[http://www.mi.ras.ru/~podolskii/files/computability1819/grading.pdf Grading]&lt;br /&gt;
&lt;br /&gt;
Send your home assignments &#039;&#039;&#039;only in pdf&#039;&#039;&#039; format to the teacher assistant (Andrey Storozhenko) by email (&#039;&#039;storozhenkoaa [at] yandex.ru&#039;&#039;) with the following subject: &amp;quot;&#039;&#039;&#039;[Computing, Name Surname, HWx]&#039;&#039;&#039;&amp;quot;. You can also submit them in person before the deadline.&lt;br /&gt;
&lt;br /&gt;
[https://docs.google.com/spreadsheets/d/1SnGKycG4m42VtKTq3Yb6rF5e0756GboLjIBO3GthpmE/edit?usp=sharing Homework results]&lt;br /&gt;
&lt;br /&gt;
== Dates and Deadlines ==&lt;br /&gt;
&lt;br /&gt;
Homework 1, deadline: 2 Oktober, before the lecture.&lt;br /&gt;
&lt;br /&gt;
== Course Materials ==&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Date !! Summary !! Problem list&lt;br /&gt;
|-&lt;br /&gt;
 || 4/9 || Time and space hierarchy theorems (see also Sipser Section 9.1) || [http://www.mi.ras.ru/~podolskii/files/computability1819/prob_1.pdf Problem list 1 ]  &lt;br /&gt;
|-&lt;br /&gt;
 || 11/9 || Complexity class NP. Examples. Inclusions between P, NP and EXP. Non-deterministic TMs. Another definition of NP. Polynomial reductions, their properties. NP-hardness and NP-completeness, their properties. || [http://www.mi.ras.ru/~podolskii/files/computability1819/prob_2.pdf Problem list 2 ]&lt;br /&gt;
|-&lt;br /&gt;
 || 18/9 || NP-completeness: Circuit-SAT, 3-SAT, NAE-3-SAT, IND-SET || [https://www.dropbox.com/s/8nisqdaia715ib0/prob_3.pdf?dl=0 Problem list 3]&lt;br /&gt;
|-&lt;br /&gt;
 || 25/9 || NP-completeness: Subset-SUM, 3COLORING || [https://www.dropbox.com/s/h5or8izfv7m5pn8/prob_4.pdf?dl=0  Problem list 4]&lt;br /&gt;
&amp;lt;!--&lt;br /&gt;
 || Space complexity. Classes PSPACE and NPSPACE. Configuration graph. Inclusions between time and space classes. TQBF problem, its PSPACE-completeness. PSPACE = NPSPACE. NSPACE(s(n)) is in SPACE(s(n)^2) (additional material). Interpretation of PSPACE in terms of games. || [http://www.mi.ras.ru/~podolskii/files/computability/prob_4.pdf Problem list 4]&lt;br /&gt;
|-&lt;br /&gt;
 || Probabilistic computation. Probabilistic machines, the class BPP, prime testing and Carmichael numbers, invariance of the definition BPP for different thresholds, RP, coRP, PP, ZPP. BPP is in P/poly.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_5.pdf Problem list 5]&lt;br /&gt;
|-&lt;br /&gt;
 || Definition of Sigma_2 and Pi_2. BPP is in Sigma_2 and Pi_2. Computations with oracles, its simple properties. There are oracles A and B such that P^A is equal to NP^A and P^B is not equal to NP^B.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_6.pdf Problem list 6] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 07.10.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Communication protocols. Functions EQ, GT, DISJ, IP. Fooling sets. Combinatorial rectangles. Rectangle size lower bound. Rank lower bound. Non-deterministic complexity. Communication complexity classes P, NP, coNP, intersection of NP and coNP.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_7.pdf Problem list 7]&lt;br /&gt;
|-&lt;br /&gt;
 || D(f)=O(N^0(f)N^1(f)). Randomized communication complexity, definitions. R(EQ)=O(1). N^1(f) vs. R^1(f). Newman&#039;s theorem, formulation.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_8.pdf Problem list 8]&lt;br /&gt;
|-&lt;br /&gt;
 || Proof of Newman&#039;s theorem, distributional complexity and the characterization of public coin communication complexity, the discrepancy method.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_9.pdf Problem list 9]&lt;br /&gt;
|-&lt;br /&gt;
 || Randomized communication complexity of IP. Streaming algorithms. Finding the majority element. Deciding whether there is a most frequent element is hard. One-sided probabilistic complexity of disjointness.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_10.pdf Problem list 10] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 22.11.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: definitions, testing of halfplanes, sorted listed, connectedness of graphs, testing of linearity.   [https://www.dropbox.com/s/r35jr22fb3lfo3k/propTest.pdf?dl=0 Lecture notes.] Version 25.11.17&lt;br /&gt;
 || [https://www.dropbox.com/s/uzh9dur6aiy88dl/prob_11.pdf?dl=0 Problem list 11] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 22.11.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: connectedness of graphs (cont.) and testing of monotonicity.  (See notes from the previous lecture.)&lt;br /&gt;
 || [https://www.dropbox.com/s/y2xujzeftsg6zlr/prob_12.pdf?dl=0 Problem list 12] &lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: lower bounds for monotonicity (see Sect 4 [http://theory.stanford.edu/~tim/w15/l/l8.pdf here]) and k-linearity using communication complexity. Approximation algorithms for some NP-complete problems (see seminar).&lt;br /&gt;
 || [https://www.dropbox.com/s/5xsxqv098y5wmx0/prob_13.pdf?dl=0 Problem list 13]&lt;br /&gt;
|-&lt;br /&gt;
 || The class PCP: definition, basic properties, and relation to with the MAX-Clique approximation problem. Beginning of the proof that NP is a subset of PCP(poly(n), 1). See Dexter Kozen, &amp;quot;Introduction to the theory of computation&amp;quot;, lectures 18-20 (or see the book of Arora and Barak, chapter 11). &lt;br /&gt;
 || [https://www.dropbox.com/s/gbg30ajwes45ne8/prob_14.pdf?dl=0 Problem list 14]&lt;br /&gt;
|- &lt;br /&gt;
 || NP is a subset of PCP(poly(n),1) [continued]. Solutions of some extra problems.&lt;br /&gt;
 || No problem list.&lt;br /&gt;
--&amp;gt;&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
During the first module, we follow Sipser&#039;s book [https://theswissbay.ch/pdf/Book/Introduction%20to%20the%20theory%20of%20computation_third%20edition%20-%20Michael%20Sipser.pdf Introduction to the theory of computation], chapters 7-9.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== Office hours ==&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! !! Person !! Monday !! Tuesday !! Wednesday !! Thursday !! Friday &lt;br /&gt;
|-&lt;br /&gt;
| &amp;lt;center&amp;gt;1&amp;lt;/center&amp;gt; || Vladimir Podolskii, room&amp;amp;nbsp;621 ||  || 18:00&amp;amp;ndash;19:00 || 16:40&amp;amp;ndash;18:00  ||  ||  &lt;br /&gt;
|-&lt;br /&gt;
| &amp;lt;center&amp;gt;2&amp;lt;/center&amp;gt; || Bruno Bauwens, room&amp;amp;nbsp;620 || 16:40&amp;amp;ndash;19:00 || 15:00&amp;amp;ndash;18:00 || ||  ||&lt;br /&gt;
|}&lt;/div&gt;</summary>
		<author><name>Storandrew</name></author>
	</entry>
	<entry>
		<id>https://wiki.cs.hse.ru/index.php?title=Theory_of_Computing_2018_2019&amp;diff=29470</id>
		<title>Theory of Computing 2018 2019</title>
		<link rel="alternate" type="text/html" href="https://wiki.cs.hse.ru/index.php?title=Theory_of_Computing_2018_2019&amp;diff=29470"/>
		<updated>2018-09-26T11:06:25Z</updated>

		<summary type="html">&lt;p&gt;Storandrew: &lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== General Information ==&lt;br /&gt;
&lt;br /&gt;
[http://www.mi.ras.ru/~podolskii/files/computability1819/grading.pdf Grading]&lt;br /&gt;
&lt;br /&gt;
Send your home assignments &#039;&#039;&#039;only in pdf&#039;&#039;&#039; format to the teacher assistant (Andrey Storozhenko) by email (&#039;&#039;storozhenkoaa [at] yandex.ru&#039;&#039;) with the following subject: &amp;quot;&#039;&#039;&#039;[Computing, Name Surname, HWx]&#039;&#039;&#039;&amp;quot;. You can also submit them in person before the deadline.&lt;br /&gt;
&lt;br /&gt;
== Dates and Deadlines ==&lt;br /&gt;
&lt;br /&gt;
Homework 1, deadline: 2 Oktober, before the lecture.&lt;br /&gt;
&lt;br /&gt;
== Course Materials ==&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Date !! Summary !! Problem list&lt;br /&gt;
|-&lt;br /&gt;
 || 4/9 || Time and space hierarchy theorems (see also Sipser Section 9.1) || [http://www.mi.ras.ru/~podolskii/files/computability1819/prob_1.pdf Problem list 1 ]  &lt;br /&gt;
|-&lt;br /&gt;
 || 11/9 || Complexity class NP. Examples. Inclusions between P, NP and EXP. Non-deterministic TMs. Another definition of NP. Polynomial reductions, their properties. NP-hardness and NP-completeness, their properties. || [http://www.mi.ras.ru/~podolskii/files/computability1819/prob_2.pdf Problem list 2 ]&lt;br /&gt;
|-&lt;br /&gt;
 || 18/9 || NP-completeness: Circuit-SAT, 3-SAT, NAE-3-SAT, IND-SET || [https://www.dropbox.com/s/8nisqdaia715ib0/prob_3.pdf?dl=0 Problem list 3]&lt;br /&gt;
|-&lt;br /&gt;
 || 25/9 || NP-completeness: Subset-SUM, 3COLORING || [https://www.dropbox.com/s/h5or8izfv7m5pn8/prob_4.pdf?dl=0  Problem list 4]&lt;br /&gt;
&amp;lt;!--&lt;br /&gt;
 || Space complexity. Classes PSPACE and NPSPACE. Configuration graph. Inclusions between time and space classes. TQBF problem, its PSPACE-completeness. PSPACE = NPSPACE. NSPACE(s(n)) is in SPACE(s(n)^2) (additional material). Interpretation of PSPACE in terms of games. || [http://www.mi.ras.ru/~podolskii/files/computability/prob_4.pdf Problem list 4]&lt;br /&gt;
|-&lt;br /&gt;
 || Probabilistic computation. Probabilistic machines, the class BPP, prime testing and Carmichael numbers, invariance of the definition BPP for different thresholds, RP, coRP, PP, ZPP. BPP is in P/poly.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_5.pdf Problem list 5]&lt;br /&gt;
|-&lt;br /&gt;
 || Definition of Sigma_2 and Pi_2. BPP is in Sigma_2 and Pi_2. Computations with oracles, its simple properties. There are oracles A and B such that P^A is equal to NP^A and P^B is not equal to NP^B.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_6.pdf Problem list 6] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 07.10.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Communication protocols. Functions EQ, GT, DISJ, IP. Fooling sets. Combinatorial rectangles. Rectangle size lower bound. Rank lower bound. Non-deterministic complexity. Communication complexity classes P, NP, coNP, intersection of NP and coNP.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_7.pdf Problem list 7]&lt;br /&gt;
|-&lt;br /&gt;
 || D(f)=O(N^0(f)N^1(f)). Randomized communication complexity, definitions. R(EQ)=O(1). N^1(f) vs. R^1(f). Newman&#039;s theorem, formulation.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_8.pdf Problem list 8]&lt;br /&gt;
|-&lt;br /&gt;
 || Proof of Newman&#039;s theorem, distributional complexity and the characterization of public coin communication complexity, the discrepancy method.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_9.pdf Problem list 9]&lt;br /&gt;
|-&lt;br /&gt;
 || Randomized communication complexity of IP. Streaming algorithms. Finding the majority element. Deciding whether there is a most frequent element is hard. One-sided probabilistic complexity of disjointness.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_10.pdf Problem list 10] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 22.11.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: definitions, testing of halfplanes, sorted listed, connectedness of graphs, testing of linearity.   [https://www.dropbox.com/s/r35jr22fb3lfo3k/propTest.pdf?dl=0 Lecture notes.] Version 25.11.17&lt;br /&gt;
 || [https://www.dropbox.com/s/uzh9dur6aiy88dl/prob_11.pdf?dl=0 Problem list 11] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 22.11.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: connectedness of graphs (cont.) and testing of monotonicity.  (See notes from the previous lecture.)&lt;br /&gt;
 || [https://www.dropbox.com/s/y2xujzeftsg6zlr/prob_12.pdf?dl=0 Problem list 12] &lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: lower bounds for monotonicity (see Sect 4 [http://theory.stanford.edu/~tim/w15/l/l8.pdf here]) and k-linearity using communication complexity. Approximation algorithms for some NP-complete problems (see seminar).&lt;br /&gt;
 || [https://www.dropbox.com/s/5xsxqv098y5wmx0/prob_13.pdf?dl=0 Problem list 13]&lt;br /&gt;
|-&lt;br /&gt;
 || The class PCP: definition, basic properties, and relation to with the MAX-Clique approximation problem. Beginning of the proof that NP is a subset of PCP(poly(n), 1). See Dexter Kozen, &amp;quot;Introduction to the theory of computation&amp;quot;, lectures 18-20 (or see the book of Arora and Barak, chapter 11). &lt;br /&gt;
 || [https://www.dropbox.com/s/gbg30ajwes45ne8/prob_14.pdf?dl=0 Problem list 14]&lt;br /&gt;
|- &lt;br /&gt;
 || NP is a subset of PCP(poly(n),1) [continued]. Solutions of some extra problems.&lt;br /&gt;
 || No problem list.&lt;br /&gt;
--&amp;gt;&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
During the first module, we follow Sipser&#039;s book [https://theswissbay.ch/pdf/Book/Introduction%20to%20the%20theory%20of%20computation_third%20edition%20-%20Michael%20Sipser.pdf Introduction to the theory of computation], chapters 7-9.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== Office hours ==&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! !! Person !! Monday !! Tuesday !! Wednesday !! Thursday !! Friday &lt;br /&gt;
|-&lt;br /&gt;
| &amp;lt;center&amp;gt;1&amp;lt;/center&amp;gt; || Vladimir Podolskii, room&amp;amp;nbsp;621 ||  || 18:00&amp;amp;ndash;19:00 || 16:40&amp;amp;ndash;18:00  ||  ||  &lt;br /&gt;
|-&lt;br /&gt;
| &amp;lt;center&amp;gt;2&amp;lt;/center&amp;gt; || Bruno Bauwens, room&amp;amp;nbsp;620 || 16:40&amp;amp;ndash;19:00 || 15:00&amp;amp;ndash;18:00 || ||  ||&lt;br /&gt;
|}&lt;/div&gt;</summary>
		<author><name>Storandrew</name></author>
	</entry>
	<entry>
		<id>https://wiki.cs.hse.ru/index.php?title=Theory_of_Computing_2018_2019&amp;diff=29469</id>
		<title>Theory of Computing 2018 2019</title>
		<link rel="alternate" type="text/html" href="https://wiki.cs.hse.ru/index.php?title=Theory_of_Computing_2018_2019&amp;diff=29469"/>
		<updated>2018-09-26T11:04:28Z</updated>

		<summary type="html">&lt;p&gt;Storandrew: /* General Information */&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== General Information ==&lt;br /&gt;
&lt;br /&gt;
[http://www.mi.ras.ru/~podolskii/files/computability1819/grading.pdf Grading]&lt;br /&gt;
&lt;br /&gt;
Send your home assignments to the teacher assistant (Andrey Storozhenko) by email (storozhenkoaa [at] yandex.ru) with the following subject: &amp;quot;[Computing, Name Surname, HWx]&amp;quot;. You can also submit them in person before the deadline.&lt;br /&gt;
&lt;br /&gt;
== Dates and Deadlines ==&lt;br /&gt;
&lt;br /&gt;
Homework 1, deadline: 2 Oktober, before the lecture.&lt;br /&gt;
&lt;br /&gt;
== Course Materials ==&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Date !! Summary !! Problem list&lt;br /&gt;
|-&lt;br /&gt;
 || 4/9 || Time and space hierarchy theorems (see also Sipser Section 9.1) || [http://www.mi.ras.ru/~podolskii/files/computability1819/prob_1.pdf Problem list 1 ]  &lt;br /&gt;
|-&lt;br /&gt;
 || 11/9 || Complexity class NP. Examples. Inclusions between P, NP and EXP. Non-deterministic TMs. Another definition of NP. Polynomial reductions, their properties. NP-hardness and NP-completeness, their properties. || [http://www.mi.ras.ru/~podolskii/files/computability1819/prob_2.pdf Problem list 2 ]&lt;br /&gt;
|-&lt;br /&gt;
 || 18/9 || NP-completeness: Circuit-SAT, 3-SAT, NAE-3-SAT, IND-SET || [https://www.dropbox.com/s/8nisqdaia715ib0/prob_3.pdf?dl=0 Problem list 3]&lt;br /&gt;
|-&lt;br /&gt;
 || 25/9 || NP-completeness: Subset-SUM, 3COLORING || [https://www.dropbox.com/s/h5or8izfv7m5pn8/prob_4.pdf?dl=0  Problem list 4]&lt;br /&gt;
&amp;lt;!--&lt;br /&gt;
 || Space complexity. Classes PSPACE and NPSPACE. Configuration graph. Inclusions between time and space classes. TQBF problem, its PSPACE-completeness. PSPACE = NPSPACE. NSPACE(s(n)) is in SPACE(s(n)^2) (additional material). Interpretation of PSPACE in terms of games. || [http://www.mi.ras.ru/~podolskii/files/computability/prob_4.pdf Problem list 4]&lt;br /&gt;
|-&lt;br /&gt;
 || Probabilistic computation. Probabilistic machines, the class BPP, prime testing and Carmichael numbers, invariance of the definition BPP for different thresholds, RP, coRP, PP, ZPP. BPP is in P/poly.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_5.pdf Problem list 5]&lt;br /&gt;
|-&lt;br /&gt;
 || Definition of Sigma_2 and Pi_2. BPP is in Sigma_2 and Pi_2. Computations with oracles, its simple properties. There are oracles A and B such that P^A is equal to NP^A and P^B is not equal to NP^B.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_6.pdf Problem list 6] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 07.10.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Communication protocols. Functions EQ, GT, DISJ, IP. Fooling sets. Combinatorial rectangles. Rectangle size lower bound. Rank lower bound. Non-deterministic complexity. Communication complexity classes P, NP, coNP, intersection of NP and coNP.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_7.pdf Problem list 7]&lt;br /&gt;
|-&lt;br /&gt;
 || D(f)=O(N^0(f)N^1(f)). Randomized communication complexity, definitions. R(EQ)=O(1). N^1(f) vs. R^1(f). Newman&#039;s theorem, formulation.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_8.pdf Problem list 8]&lt;br /&gt;
|-&lt;br /&gt;
 || Proof of Newman&#039;s theorem, distributional complexity and the characterization of public coin communication complexity, the discrepancy method.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_9.pdf Problem list 9]&lt;br /&gt;
|-&lt;br /&gt;
 || Randomized communication complexity of IP. Streaming algorithms. Finding the majority element. Deciding whether there is a most frequent element is hard. One-sided probabilistic complexity of disjointness.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_10.pdf Problem list 10] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 22.11.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: definitions, testing of halfplanes, sorted listed, connectedness of graphs, testing of linearity.   [https://www.dropbox.com/s/r35jr22fb3lfo3k/propTest.pdf?dl=0 Lecture notes.] Version 25.11.17&lt;br /&gt;
 || [https://www.dropbox.com/s/uzh9dur6aiy88dl/prob_11.pdf?dl=0 Problem list 11] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 22.11.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: connectedness of graphs (cont.) and testing of monotonicity.  (See notes from the previous lecture.)&lt;br /&gt;
 || [https://www.dropbox.com/s/y2xujzeftsg6zlr/prob_12.pdf?dl=0 Problem list 12] &lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: lower bounds for monotonicity (see Sect 4 [http://theory.stanford.edu/~tim/w15/l/l8.pdf here]) and k-linearity using communication complexity. Approximation algorithms for some NP-complete problems (see seminar).&lt;br /&gt;
 || [https://www.dropbox.com/s/5xsxqv098y5wmx0/prob_13.pdf?dl=0 Problem list 13]&lt;br /&gt;
|-&lt;br /&gt;
 || The class PCP: definition, basic properties, and relation to with the MAX-Clique approximation problem. Beginning of the proof that NP is a subset of PCP(poly(n), 1). See Dexter Kozen, &amp;quot;Introduction to the theory of computation&amp;quot;, lectures 18-20 (or see the book of Arora and Barak, chapter 11). &lt;br /&gt;
 || [https://www.dropbox.com/s/gbg30ajwes45ne8/prob_14.pdf?dl=0 Problem list 14]&lt;br /&gt;
|- &lt;br /&gt;
 || NP is a subset of PCP(poly(n),1) [continued]. Solutions of some extra problems.&lt;br /&gt;
 || No problem list.&lt;br /&gt;
--&amp;gt;&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
During the first module, we follow Sipser&#039;s book [https://theswissbay.ch/pdf/Book/Introduction%20to%20the%20theory%20of%20computation_third%20edition%20-%20Michael%20Sipser.pdf Introduction to the theory of computation], chapters 7-9.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== Office hours ==&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! !! Person !! Monday !! Tuesday !! Wednesday !! Thursday !! Friday &lt;br /&gt;
|-&lt;br /&gt;
| &amp;lt;center&amp;gt;1&amp;lt;/center&amp;gt; || Vladimir Podolskii, room&amp;amp;nbsp;621 ||  || 18:00&amp;amp;ndash;19:00 || 16:40&amp;amp;ndash;18:00  ||  ||  &lt;br /&gt;
|-&lt;br /&gt;
| &amp;lt;center&amp;gt;2&amp;lt;/center&amp;gt; || Bruno Bauwens, room&amp;amp;nbsp;620 || 16:40&amp;amp;ndash;19:00 || 15:00&amp;amp;ndash;18:00 || ||  ||&lt;br /&gt;
|}&lt;/div&gt;</summary>
		<author><name>Storandrew</name></author>
	</entry>
	<entry>
		<id>https://wiki.cs.hse.ru/index.php?title=Theory_of_Computing_2018_2019&amp;diff=29464</id>
		<title>Theory of Computing 2018 2019</title>
		<link rel="alternate" type="text/html" href="https://wiki.cs.hse.ru/index.php?title=Theory_of_Computing_2018_2019&amp;diff=29464"/>
		<updated>2018-09-26T09:40:48Z</updated>

		<summary type="html">&lt;p&gt;Storandrew: &lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== General Information ==&lt;br /&gt;
&lt;br /&gt;
[http://www.mi.ras.ru/~podolskii/files/computability1819/grading.pdf Grading]&lt;br /&gt;
&lt;br /&gt;
Send your home assignment to the teacher assistant (Andrey Storozhenko) by email (storozhenkoaa [at] yandex.ru) or submit them in person before the deadline.&lt;br /&gt;
&lt;br /&gt;
== Dates and Deadlines ==&lt;br /&gt;
&lt;br /&gt;
Homework 1, deadline: 2 Oktober, before the lecture.&lt;br /&gt;
&lt;br /&gt;
== Course Materials ==&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Date !! Summary !! Problem list&lt;br /&gt;
|-&lt;br /&gt;
 || 4/9 || Time and space hierarchy theorems (see also Sipser Section 9.1) || [http://www.mi.ras.ru/~podolskii/files/computability1819/prob_1.pdf Problem list 1 ]  &lt;br /&gt;
|-&lt;br /&gt;
 || 11/9 || Complexity class NP. Examples. Inclusions between P, NP and EXP. Non-deterministic TMs. Another definition of NP. Polynomial reductions, their properties. NP-hardness and NP-completeness, their properties. || [http://www.mi.ras.ru/~podolskii/files/computability1819/prob_2.pdf Problem list 2 ]&lt;br /&gt;
|-&lt;br /&gt;
 || 18/9 || NP-completeness: Circuit-SAT, 3-SAT, NAE-3-SAT, IND-SET || [https://www.dropbox.com/s/8nisqdaia715ib0/prob_3.pdf?dl=0 Problem list 3]&lt;br /&gt;
|-&lt;br /&gt;
 || 25/9 || NP-completeness: Subset-SUM, 3COLORING || [https://www.dropbox.com/s/h5or8izfv7m5pn8/prob_4.pdf?dl=0  Problem list 4]&lt;br /&gt;
&amp;lt;!--&lt;br /&gt;
 || Space complexity. Classes PSPACE and NPSPACE. Configuration graph. Inclusions between time and space classes. TQBF problem, its PSPACE-completeness. PSPACE = NPSPACE. NSPACE(s(n)) is in SPACE(s(n)^2) (additional material). Interpretation of PSPACE in terms of games. || [http://www.mi.ras.ru/~podolskii/files/computability/prob_4.pdf Problem list 4]&lt;br /&gt;
|-&lt;br /&gt;
 || Probabilistic computation. Probabilistic machines, the class BPP, prime testing and Carmichael numbers, invariance of the definition BPP for different thresholds, RP, coRP, PP, ZPP. BPP is in P/poly.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_5.pdf Problem list 5]&lt;br /&gt;
|-&lt;br /&gt;
 || Definition of Sigma_2 and Pi_2. BPP is in Sigma_2 and Pi_2. Computations with oracles, its simple properties. There are oracles A and B such that P^A is equal to NP^A and P^B is not equal to NP^B.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_6.pdf Problem list 6] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 07.10.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Communication protocols. Functions EQ, GT, DISJ, IP. Fooling sets. Combinatorial rectangles. Rectangle size lower bound. Rank lower bound. Non-deterministic complexity. Communication complexity classes P, NP, coNP, intersection of NP and coNP.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_7.pdf Problem list 7]&lt;br /&gt;
|-&lt;br /&gt;
 || D(f)=O(N^0(f)N^1(f)). Randomized communication complexity, definitions. R(EQ)=O(1). N^1(f) vs. R^1(f). Newman&#039;s theorem, formulation.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_8.pdf Problem list 8]&lt;br /&gt;
|-&lt;br /&gt;
 || Proof of Newman&#039;s theorem, distributional complexity and the characterization of public coin communication complexity, the discrepancy method.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_9.pdf Problem list 9]&lt;br /&gt;
|-&lt;br /&gt;
 || Randomized communication complexity of IP. Streaming algorithms. Finding the majority element. Deciding whether there is a most frequent element is hard. One-sided probabilistic complexity of disjointness.&lt;br /&gt;
 || [http://www.mi.ras.ru/~podolskii/files/computability/prob_10.pdf Problem list 10] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 22.11.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: definitions, testing of halfplanes, sorted listed, connectedness of graphs, testing of linearity.   [https://www.dropbox.com/s/r35jr22fb3lfo3k/propTest.pdf?dl=0 Lecture notes.] Version 25.11.17&lt;br /&gt;
 || [https://www.dropbox.com/s/uzh9dur6aiy88dl/prob_11.pdf?dl=0 Problem list 11] &amp;lt;span style=&amp;quot;color:red&amp;quot;&amp;gt;Updated: 22.11.17&amp;lt;/span&amp;gt;&lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: connectedness of graphs (cont.) and testing of monotonicity.  (See notes from the previous lecture.)&lt;br /&gt;
 || [https://www.dropbox.com/s/y2xujzeftsg6zlr/prob_12.pdf?dl=0 Problem list 12] &lt;br /&gt;
|-&lt;br /&gt;
 || Property testing: lower bounds for monotonicity (see Sect 4 [http://theory.stanford.edu/~tim/w15/l/l8.pdf here]) and k-linearity using communication complexity. Approximation algorithms for some NP-complete problems (see seminar).&lt;br /&gt;
 || [https://www.dropbox.com/s/5xsxqv098y5wmx0/prob_13.pdf?dl=0 Problem list 13]&lt;br /&gt;
|-&lt;br /&gt;
 || The class PCP: definition, basic properties, and relation to with the MAX-Clique approximation problem. Beginning of the proof that NP is a subset of PCP(poly(n), 1). See Dexter Kozen, &amp;quot;Introduction to the theory of computation&amp;quot;, lectures 18-20 (or see the book of Arora and Barak, chapter 11). &lt;br /&gt;
 || [https://www.dropbox.com/s/gbg30ajwes45ne8/prob_14.pdf?dl=0 Problem list 14]&lt;br /&gt;
|- &lt;br /&gt;
 || NP is a subset of PCP(poly(n),1) [continued]. Solutions of some extra problems.&lt;br /&gt;
 || No problem list.&lt;br /&gt;
--&amp;gt;&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
During the first module, we follow Sipser&#039;s book [https://theswissbay.ch/pdf/Book/Introduction%20to%20the%20theory%20of%20computation_third%20edition%20-%20Michael%20Sipser.pdf Introduction to the theory of computation], chapters 7-9.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== Office hours ==&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! !! Person !! Monday !! Tuesday !! Wednesday !! Thursday !! Friday &lt;br /&gt;
|-&lt;br /&gt;
| &amp;lt;center&amp;gt;1&amp;lt;/center&amp;gt; || Vladimir Podolskii, room&amp;amp;nbsp;621 ||  || 18:00&amp;amp;ndash;19:00 || 16:40&amp;amp;ndash;18:00  ||  ||  &lt;br /&gt;
|-&lt;br /&gt;
| &amp;lt;center&amp;gt;2&amp;lt;/center&amp;gt; || Bruno Bauwens, room&amp;amp;nbsp;620 || 16:40&amp;amp;ndash;19:00 || 15:00&amp;amp;ndash;18:00 || ||  ||&lt;br /&gt;
|}&lt;/div&gt;</summary>
		<author><name>Storandrew</name></author>
	</entry>
	<entry>
		<id>https://wiki.cs.hse.ru/index.php?title=%D0%A0%D0%B5%D0%BD%D0%B4%D0%B7%D1%8E_(%D1%81%D0%B5%D0%BC%D0%B8%D0%BD%D0%B0%D1%80)&amp;diff=22489</id>
		<title>Рендзю (семинар)</title>
		<link rel="alternate" type="text/html" href="https://wiki.cs.hse.ru/index.php?title=%D0%A0%D0%B5%D0%BD%D0%B4%D0%B7%D1%8E_(%D1%81%D0%B5%D0%BC%D0%B8%D0%BD%D0%B0%D1%80)&amp;diff=22489"/>
		<updated>2017-01-28T13:10:11Z</updated>

		<summary type="html">&lt;p&gt;Storandrew: &lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;Описание [[Рендзю (проект)|проекта]], последний [[Рендзю_(семинар)#.D0.A1.D0.B5.D0.BC.D0.B8.D0.BD.D0.B0.D1.80.D1.8B|семинар]].&lt;br /&gt;
&lt;br /&gt;
==Правила игры ==&lt;br /&gt;
* &#039;&#039;&#039;Ментор:&#039;&#039;&#039; [[Участник:Simagin.denis|Симагин Денис]].&lt;br /&gt;
* &#039;&#039;&#039;Место:&#039;&#039;&#039; офис Яндекса ([https://maps.yandex.ru/213/moscow/?ll=37.590150%2C55.734065&amp;amp;z=18&amp;amp;l=stv%2Csta&amp;amp;panorama%5Bpoint%5D=37.589416%2C55.733747&amp;amp;panorama%5Bdirection%5D=40.412258%2C-11.910596&amp;amp;panorama%5Bspan%5D=130.000000%2C52.209677 место встречи])&lt;br /&gt;
* &#039;&#039;&#039;Время:&#039;&#039;&#039; c 19:00, каждую среду.&lt;br /&gt;
&lt;br /&gt;
Общение с ментором вне занятий приветствуется. Можно задавать вопросы, в том числе философские. Но перед тем, как написать, попробуйте спросить это у [https://ya.ru Яндекса]. Также не обижайтесь, если в ответ вам пришла ссылка на документацию или какую-то статью.&lt;br /&gt;
&lt;br /&gt;
===Ключевые точки===&lt;br /&gt;
Сверху нам спущены ключевые точки выполнения проекта. Для нас они скорее явлются формальными, тем не менее мы должны их соблюдать.&lt;br /&gt;
# &#039;&#039;&#039;12-17 декабря&#039;&#039;&#039; - все включились в работу&lt;br /&gt;
# &#039;&#039;&#039;20-25 марта&#039;&#039;&#039; - реализован объем работ, необходимый для зачета&lt;br /&gt;
# &#039;&#039;&#039;30 мая - 3 июня&#039;&#039;&#039; - окончание проектной работы, вы готовы, как пионеры. &lt;br /&gt;
# &#039;&#039;&#039;начало июня&#039;&#039;&#039; - конкурс проектов.&lt;br /&gt;
&lt;br /&gt;
===Правило 2Х===&lt;br /&gt;
У вас есть право на одну ошибку. Следующая - я отказываюсь с вами работать.&lt;br /&gt;
&lt;br /&gt;
===Репозитории===&lt;br /&gt;
Студенты хранят свой код в следующих репозиториях&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
| Ментор  || https://github.com/dasimagin/renju&lt;br /&gt;
|-&lt;br /&gt;
| Харламов || https://github.com/gamers5a/renju&lt;br /&gt;
|-&lt;br /&gt;
| Сопов || https://github.com/PreFX48/renju&lt;br /&gt;
|-&lt;br /&gt;
| Vodim  || https://github.com/EterniusVGM/Renju&lt;br /&gt;
|-&lt;br /&gt;
| Yuriy || https://github.com/yurriy/renju&lt;br /&gt;
|-&lt;br /&gt;
| Storozh || https://github.com/storandrew/Renju&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
===Разбор статьи===&lt;br /&gt;
В рамках проекта студент должен разобрать интересную для него статью и доложить ее на общем семинаре.&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Студент !! Статья !! Дата&lt;br /&gt;
|-&lt;br /&gt;
| Харламов || http://web.stanford.edu/~takapoui/linear_bandits.pdf || -&lt;br /&gt;
|-&lt;br /&gt;
| Сопов || https://papers.nips.cc/paper/6068-learning-feed-forward-one-shot-learners.pdf || -&lt;br /&gt;
|-&lt;br /&gt;
| Гринберг  || https://arxiv.org/pdf/1611.01626.pdf || -&lt;br /&gt;
|-&lt;br /&gt;
| Баранов || https://arxiv.org/pdf/1611.01224.pdf || -&lt;br /&gt;
|-&lt;br /&gt;
| Стороженко || https://arxiv.org/pdf/1511.06581v3.pdf || -&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
===Лабораторные===&lt;br /&gt;
Лабораторные проводятся для практического закрепления материала. Их выполнение учитывается в итоговой оценке.&lt;br /&gt;
&lt;br /&gt;
# Результатом работы является jupyter notebook, где сохранен вывод вашего кода, графики и т.п. А так же его импорт в формат .py. Для автоматизации процесса можно настроить jupyter.&lt;br /&gt;
# Когда сроки выполнения лабораторной завершены, вы выкладываете ее на ревью, создавая соответствующее задание и запрос на объединение ветки с мастером (не забудьте добавить проверяющего).&lt;br /&gt;
# Ваш коллега проводит ревью кода и может оставлять замечания, как в виде комментариев к заданию, так и в файле .py. Оно предполагает проверку стиля и правильность кода, а также конструктивные замечания по производительности. Однако не стремитесь сразу оптимизировать код. Добейтесь лучше того, чтобы все работало правильно.&lt;br /&gt;
# Когда ревью завершено, влейтесь в мастер и закройте задание.&lt;br /&gt;
&lt;br /&gt;
===Результаты===&lt;br /&gt;
Текущие результаты можно найти  [https://docs.google.com/spreadsheets/d/1VAaIoKGOYkMYKxYMPjHs_TsYabjHo27jA4fiQg_R344/edit?usp=sharing здесь]. Оценка складывается из нескольких частей:&lt;br /&gt;
# Работа на семинаре&lt;br /&gt;
# Доклад статьи&lt;br /&gt;
# Итоговое качество игры&lt;br /&gt;
&lt;br /&gt;
==Семинары==&lt;br /&gt;
&lt;br /&gt;
===S01.02, S08.02, 15.02, 22.02, 01.03===&lt;br /&gt;
Разбор статьи &lt;br /&gt;
===S25.01===&lt;br /&gt;
# Поговорили на тему [https://en.wikipedia.org/wiki/Multi-armed_bandit многоруких бандитов].&lt;br /&gt;
# Начали осваивать [https://en.wikipedia.org/wiki/Reinforcement_learning reinforcment learning].&lt;br /&gt;
&lt;br /&gt;
Полезная книга может быть найдена [https://webdocs.cs.ualberta.ca/~sutton/book/bookdraft2016sep.pdf  здесь].&lt;br /&gt;
&lt;br /&gt;
===L4===&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Ревьюер !! Разработчик !! Оценка&lt;br /&gt;
|-&lt;br /&gt;
| Сопов || Харламов|| -&lt;br /&gt;
|-&lt;br /&gt;
| Гринберг || Сопов  || -&lt;br /&gt;
|-&lt;br /&gt;
| Баранов  || Гринберг || -&lt;br /&gt;
|-&lt;br /&gt;
| Стороженко || Баранов || -&lt;br /&gt;
|-&lt;br /&gt;
| Харламов  || Стороженко || -&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
Появилась очередная лабораторная работа [https://github.com/dasimagin/renju/blob/master/labs/L4%20-%20Reinforcement%20learning.ipynb L4], сроки &#039;&#039;&#039;...&#039;&#039;&#039;.&lt;br /&gt;
&lt;br /&gt;
===S18.01===&lt;br /&gt;
В связи с болезнью ментора занятие отменено.&lt;br /&gt;
&lt;br /&gt;
===H11.01===&lt;br /&gt;
Произвели разбор L3. Описание модели победителя можно найти [здесь], а baseline доступен [https://github.com/dasimagin/renju/blob/master/labs/L3%20-%20Baseline.ipynb здесь].&lt;br /&gt;
&lt;br /&gt;
Доклады мне не очень понравились. Постараюсь написать общие замечания.&lt;br /&gt;
# Прежде всего у докладчика должна быть хорошая речь.&lt;br /&gt;
# Нужно выделить то, что действительно важно и интересно для слушателя.&lt;br /&gt;
# Делать на доске четкие и простые рисунки и записи, убедиться, что аудитория тебя понимает.&lt;br /&gt;
# Не прыгать с темы на тему, а идти в соответсвии с логическим планом.&lt;br /&gt;
&lt;br /&gt;
===S14.12===&lt;br /&gt;
Начали разбирать нашу [https://storage.googleapis.com/deepmind-media/alphago/AlphaGoNaturePaper.pdf статью]. Есть пара источников на русском:&lt;br /&gt;
* [https://ru.wikipedia.org/wiki/AlphaGo Статья] на wikipedia&lt;br /&gt;
* [https://habrahabr.ru/post/279071/ Статья]  на хабре&lt;br /&gt;
&lt;br /&gt;
===H11.12===&lt;br /&gt;
Занятие было посвящено выполнению второй лабораторной. Интересный ноутбук скоро появится [здесь].&lt;br /&gt;
&lt;br /&gt;
===S08.12===&lt;br /&gt;
&#039;&#039;&#039;1. Известные архитектуры сверточных сетей&#039;&#039;&#039;&lt;br /&gt;
* [https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf Alexnet]&lt;br /&gt;
* [https://arxiv.org/pdf/1409.1556.pdf VGG net]&lt;br /&gt;
* [https://arxiv.org/pdf/1409.4842v1.pdf GoogLeNet]&lt;br /&gt;
* [https://arxiv.org/pdf/1512.03385.pdf ResNet]&lt;br /&gt;
&#039;&#039;&#039;2. Поговорили:&#039;&#039;&#039;&lt;br /&gt;
* На что активируются нейроны в зависимости от слоя&lt;br /&gt;
* Генерация &#039;похожих картинок&#039;&lt;br /&gt;
* Послойное обучение сети&lt;br /&gt;
* Переобучение или дообучение уже готовой сети&lt;br /&gt;
&#039;&#039;&#039;3. Изучили примеры для библиотеки Keras&#039;&#039;&#039;&lt;br /&gt;
*  [https://github.com/fchollet/keras/blob/master/examples/mnist_mlp.py Полносвязанная сеть]&lt;br /&gt;
* [https://github.com/fchollet/keras/blob/master/examples/mnist_cnn.py Сверточная сеть ]&lt;br /&gt;
* [https://github.com/fchollet/keras/blob/master/examples/mnist_transfer_cnn.py Переобучение] сверточной сети &lt;br /&gt;
&lt;br /&gt;
===L3===&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Студент !! CPU !! RAM !! GPU&lt;br /&gt;
|- &lt;br /&gt;
| Пример || 6 core, 3,5 GHz || 64GB || NVIDIA TITAN X&lt;br /&gt;
|-&lt;br /&gt;
| Харламов || 4 core, 3,6 GHz || 16GB || NVIDIA GTX 960m 2GB&lt;br /&gt;
|-&lt;br /&gt;
| Сопов || 4 core, 2,7 GHz || 8GB || NVIDIA GTX 940m&lt;br /&gt;
|-&lt;br /&gt;
| Гринберг || 4 core, 3,6 GHz || 16 GB || NVIDIA GTX 1070 8GB&lt;br /&gt;
|-&lt;br /&gt;
| Баранов || 4 core, 3,5GHz || 8GB || NVIDIA GTX 960m 2GB&lt;br /&gt;
|-&lt;br /&gt;
| Стороженко || 6 core 3.0 GHz || 16 GB || NVIDIA GTX 1060 6GB&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
Победить в [https://inclass.kaggle.com/c/ch-ch конкурсе] классификации,  срок 3 января, 23:59.&lt;br /&gt;
&lt;br /&gt;
Для этого вам понадобится&lt;br /&gt;
* Установить [https://www.tensorflow.org Tensorflow] &lt;br /&gt;
* Установить [https://keras.io Keras] &lt;br /&gt;
* Запастись терпением&lt;br /&gt;
&lt;br /&gt;
===S01.12===&lt;br /&gt;
&#039;&#039;&#039;1. Полносвязанные сети:&#039;&#039;&#039;&lt;br /&gt;
* Подсчитаны производные для [https://en.wikipedia.org/wiki/Backpropagation Backpropagation], обсуждены тонкости реализации.&lt;br /&gt;
* Различные виды нелинейности: [https://en.wikipedia.org/wiki/Rectifier_(neural_networks) ReLu], [https://arxiv.org/pdf/1502.01852v1.pdf PReLu], [https://en.wikipedia.org/wiki/Sigmoid_function Sigmoid].&lt;br /&gt;
* Обучение сетей при помощи [https://en.wikipedia.org/wiki/Autoencoder Autoencoder].&lt;br /&gt;
&lt;br /&gt;
&#039;&#039;&#039;2. Сверточные сети:&#039;&#039;&#039;&lt;br /&gt;
* Cтруктура [https://en.wikipedia.org/wiki/Convolutional_neural_network CNN].&lt;br /&gt;
* Затронуты: [https://en.wikipedia.org/wiki/Convolution Convolution], [https://en.wikipedia.org/wiki/Convolutional_neural_network#Pooling_layer Pooling].&lt;br /&gt;
* Влияние различных ядер свертки на структуру сети.&lt;br /&gt;
* [https://en.wikipedia.org/wiki/Convolutional_neural_network#Choosing_hyperparameters Feature maps].&lt;br /&gt;
* Разобрана архитектура [https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf Alexnet].&lt;br /&gt;
* Сочетание из Convolutional и Dense слоев.&lt;br /&gt;
&lt;br /&gt;
===S24.11===&lt;br /&gt;
&#039;&#039;&#039;1. Регуляризация:&#039;&#039;&#039;&lt;br /&gt;
* Разобрали L1 и L2 регуляризаторы, можно найти [https://en.wikipedia.org/wiki/Regularization_(mathematics) здесь].&lt;br /&gt;
* Используйте простые классификаторы&lt;br /&gt;
* Раняя остановка (смотрим качество на отложенном множестве)&lt;br /&gt;
* Добавление шума&lt;br /&gt;
* Комбинирование классификаторов&lt;br /&gt;
&lt;br /&gt;
&#039;&#039;&#039;2. Полносвязанные сети:&#039;&#039;&#039;&lt;br /&gt;
* Множественная классификация и [https://en.wikipedia.org/wiki/Softmax_function softmax].&lt;br /&gt;
* Метод [http://www.machinelearning.ru/wiki/images/0/0f/karpinskaya-2010.pdf обратного распространения ошибки], проблема при обучении.&lt;br /&gt;
* Инициализация весов:  и [http://jmlr.org/proceedings/papers/v9/glorot10a/glorot10a.pdf xavier] и другие [https://arxiv.org/pdf/1502.01852v1.pdf вариации].&lt;br /&gt;
* Кратко о [https://en.wikipedia.org/wiki/Convolutional_neural_network#Dropout dropout].&lt;br /&gt;
&lt;br /&gt;
===L2===&lt;br /&gt;
Задание можно найти [https://github.com/dasimagin/renju/blob/master/labs/L2%20-%20Nets.ipynb здесь],  срок 23:59 11 декабря.&lt;br /&gt;
&lt;br /&gt;
===S02.11===&lt;br /&gt;
# [https://en.wikipedia.org/wiki/Feature_(machine_learning) Признаки] и какие они бывают. Об отборе признаков, кратко [https://habrahabr.ru/post/264915/ тут]. Может помочь на конкурсе.&lt;br /&gt;
# Задача [http://www.machinelearning.ru/wiki/index.php?title=Линейный_классификатор бинарной классификации].&lt;br /&gt;
# [http://www.machinelearning.ru/wiki/index.php?title=Метод_градиентного_спуска Градиентный спуск].&lt;br /&gt;
# [http://www.machinelearning.ru/wiki/index.php?title=Метод_стохастического_градиента Стохастический градиентный спуск]. На английской [https://en.wikipedia.org/wiki/Stochastic_gradient_descent вике] больше интересной информации.&lt;br /&gt;
&lt;br /&gt;
&#039;&#039;&#039;Для дополнительного чтения:&#039;&#039;&#039;&lt;br /&gt;
# [https://homes.cs.washington.edu/~pedrod/papers/cacm12.pdf Что полезно знать о машинном обучении].&lt;br /&gt;
# [https://en.wikipedia.org/wiki/Feature_(machine_learning) Английская вика про признаки]&lt;br /&gt;
# [http://www.jmlr.org/papers/volume3/guyon03a/guyon03a.pdf Отбор признаков].&lt;br /&gt;
# Мощная теоретическая работа про [https://mipt.ru/upload/medialibrary/d7e/41-91.pdf стохастический градиентный спуск].&lt;br /&gt;
&lt;br /&gt;
===L1===&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Ревьюер !! Разработчик !! Оценка&lt;br /&gt;
|-&lt;br /&gt;
|Харламов || Сопов || 9&lt;br /&gt;
|-&lt;br /&gt;
| Сопов || Гринберг || 10&lt;br /&gt;
|-&lt;br /&gt;
| Гринберг || Баранов || 8&lt;br /&gt;
|-&lt;br /&gt;
| Баранов || Стороженко || 10&lt;br /&gt;
|-&lt;br /&gt;
| Стороженко || Харламов || 8&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
Для [https://github.com/dasimagin/renju/blob/master/labs/L1%20-%20Gradient%20descent%20and%20linear%20models.ipynb первой] лабораторной работы вам потребуется:&lt;br /&gt;
# Настроить себе [https://pip.pypa.io/en/stable/ pip] для Python3&lt;br /&gt;
# Освоить [http://jupyter.org Jupyter notebook]&lt;br /&gt;
# Установить пакеты [http://www.scipy.org scipy]: numpy, scipy, matplotlib&lt;/div&gt;</summary>
		<author><name>Storandrew</name></author>
	</entry>
</feed>