deep machine learning as a way to build a portfolio of ...ecosystem.education/doc/1deep machine...
TRANSCRIPT
1
Aleksandr Sheremetev
Information and Statistics Analyst
Moscow Technical University of Communications and Informatics
Deep machine learning as a way to build a portfolio of missing knowledge
2
Part 1
Machine learning
Machine Learning is an extensivesubsection of artificial intelligencethat studies methods for constructinglearning-capable algorithms. Thereare two types of training. Case-lawtraining, or inductive training, is basedon the identification of generalpatterns from particular empiricaldata. Deductive learning involvesformalizing expert knowledge andtransferring it to a computer in theform of a knowledge base. Deductivelearning is usually referred to the fieldof expert systems, so the termsmachine learning and case studies canbe considered synonymous.
3
Deep machine learning
The deep learning process includes the following steps:д
ANNs ask a set of binary questions in the form of yes / no.
Retrieving numeric values from data blocks.
Classification of data according to responses received.
Data labeling4
Machine learning in the educational process
25,40%
32,60%37,10%
44,00%
52,60%
57,10%
67,50%70,40% 70,40%
0,00%
10,00%
20,00%
30,00%
40,00%
50,00%
60,00%
70,00%
80,00%
2008 2009 2010 2011 2012 2013 2014 2015 2016
The audience of Internet users in Russia(www.gfk.com)
5
6
«Use of machine learning techniques for educational proposes: a decision support system for forecasting students’ grades»
China
7
Already, an intelligent system has been installed in Chinese schools, which, using deep machine learning technology,determines the knowledge map for each student. The principle of operation of such a system is that it collects data ofcompleted homework, written tests.
8
Part 2
Legislation
The Bologna Process
Federal Law No. 273-FL“About Education in theRussian Federation”
The methodology of theuse of distanceeducational technologies(distance learning) ineducational institutionsof higher, secondary andadditional professionaleducation of the RussianFederation.
9
Intelligent Education Systems
ITS models usually consist of the following component:j
• domain model;
• student model;
• learning model;
• learning environment.
10
Advantages:
In the ITS system, a large number of students can work simultaneously;
System availability at any time;
The operating mode and time spent on material for each student are individual;
The availability of the necessary material in one place;
An adjusted learning path, focused on the individual characteristics of the student;
Improving grades and understanding after working with the system.
Disadvantages:
ITS requires a lot of time for their development, selection of the necessary information and structuring of the material;
If the system provides for the conclusion of the correct answer at the end of training, then the student may lose motivation to study the material in order to get the correct answer to the question;
The effectiveness of using ITS systems begins to decline after a year of use;
Integration into the educational process as an additional type of activity.
Hight School of Economics
11
2012 2016 2012 2016
University 20.35
• WHAT IS UNIVERSITY 20.35
• University 20.35 is the first university in Russia providing opportunities for professional development by creating individualeducational trajectories and tracking digital skill profiles.
• It is aimed at training business leaders, participants of the National Technology Initiative and professionals entering new globalmarkets.University 20.35 pioneers a new network-based learning principle, where educational trajectories for each student areselected in a personalized manner. Different universities, online education platforms and other organizations are providing thiscustomized content. Students are trained both offline and online through the digital platform of the University.
• University 20.35 was established by the Agency for Strategic Initiatives (ASI) to promote new projects. Partners: Skoltech, Innopolis,ITMO, SPbPU, MIPT, Novosibirsk State University, Tomsk State University, Far Eastern Federal University.
12
MTUCI
Electronic educational information environment
Internet
Schedule
Student's personal account
Administration
E-learning resources
Personal account of the teacher
Digital library
13
14
MTUCI
15
Start self-study
Testing on a selected
topic
The topic is mastered
Go to the next topic
All tasks solved
The topic has not been mastered
Error Type Definition
There is at least one mistake
Selection of required
definitions
Definition Block Output
The error in the theoretical part
Student choice
Error in the practical part
Collection of statistical responses
Print recommenda
tions
Additional tasks
Print a ready-made solution to a problem
The topic is not learned. Additional tests will be
conducted.
MTUCI
Conclusions
Given the rapid increasein information, themassive use of tablets,gadgets, smartphonesand laptops, as well asincreasing computerliteracy of the population,the development ofintelligent educationsystems and theirintroduction into theeducational process iscritically relevant at themoment.
16
Tehran
Moscow
Thank you for attention!
17
www.mtuci.ru
facebook.com/mtuci.ru/+7 (495) 957-78-75
[email protected] mtuci.live
vk.com/mtuci
TVMTUCI Aleksandr Sheremetev
Information and Statistics Analyst
Moscow Technical University of Communications and Informatics
Find out more on: