Курсы английского
<<  The IBM view on storage archive solutions: requirements to solve and trends for the future 31st METHODS FOR DETECTION AND REMOVAL OF PARASITIC FREQUENCY MODULATION IN AUDIO RECORDINGS  >>
The future perspectives of applications of artificial neural networks
The future perspectives of applications of artificial neural networks
Outline
Outline
Conclusions
Conclusions
Conclusions
Conclusions
Conclusions
Conclusions
Introduction
Introduction
The goal
The goal
Introduction (main
Introduction (main
Data Base
Data Base
Data Analysis
Data Analysis
Ex-1: Marriage and Dissolving of Marriage
Ex-1: Marriage and Dissolving of Marriage
Ex-1: Marriage and Dissolving of Marriage
Ex-1: Marriage and Dissolving of Marriage
Ex-1
Ex-1
Ex-1
Ex-1
Ex-1
Ex-1
Ex-1
Ex-1
Ex-1
Ex-1
Ex-2: BasegroupLab Development
Ex-2: BasegroupLab Development
Ex - 2
Ex - 2
Future
Future
Future
Future
Future
Future
Future
Future
Future
Future
Future
Future
Future
Future
Conclusions
Conclusions
R&D Projects - 1
R&D Projects - 1
R&D Projects - 2
R&D Projects - 2
R&D Projects - 2
R&D Projects - 2
The first results of R&D - 2
The first results of R&D - 2
Other R&D Projects
Other R&D Projects
Other R&D Projects
Other R&D Projects
Other R&D Projects
Other R&D Projects
Other R&D Projects
Other R&D Projects
Now your practice
Now your practice
Conclusion
Conclusion

Презентация на тему: «The future perspectives of applications of artificial neural networks for diagnostics, modeling, and control in science and industry (methods, results, demonstrations hands-on)». Автор: Евгений. Файл: «The future perspectives of applications of artificial neural networks for diagnostics, modeling, and control in science and industry (methods, results, demonstrations hands-on).ppt». Размер zip-архива: 701 КБ.

The future perspectives of applications of artificial neural networks for diagnostics, modeling, and control in science and industry (methods, results, demonstrations hands-on)

содержание презентации «The future perspectives of applications of artificial neural networks for diagnostics, modeling, and control in science and industry (methods, results, demonstrations hands-on).ppt»
СлайдТекст
1 The future perspectives of applications of artificial neural networks

The future perspectives of applications of artificial neural networks

for diagnostics, modeling, and control in science and industry (methods, results, demonstrations & hands-on)

Chuvash State University, Russia Department of Thermo Physics

Victor S. Abrukov (2008) e-mail: abrukov@yandex.ru

2 Outline

Outline

1. Conclusions!? What do we know? 2. Introduction to the 3d-Seminar. Ex-1: Marriage and Dissolving of Marriage (with support of Russian Foundation for Basic Research, grant 07-06-00277) Ex-2: Tasks that have been solved by BasegroupLab Ex-3: Future. Educational Systems Ex-4: Future. Terrorism Modeling 3. Conclusions!! R&D Projects

3 Conclusions

Conclusions

? What do we know?

ANN can solve the inverse and direct problem easy and fast in spite of incompleteness of your data. ANN can solve problems of modeling, recognition, predictions, and control in science and engineering.

4 Conclusions

Conclusions

? What you must know!

There are examples of many other ANN applications: condition monitoring, fault diagnosis, industrial control, medical and diagnostic systems; image processing, adaptive filtration and recognition; smart machine and computer vision, authentification systems, environmental monitoring, power electronics and drives, high voltage systems, electronic noses, engine control and vehicle applications, "smart Landers", searching of specific attributes and elimination of noise in complex electrical signals, engine control, tracking system of a state of equipment, intelligent flight control system, solving of adaptive nonlinear filtering of aircraft movement parameters for processing the radar-tracking information at a knot, detection and support of movement trajectories, control scheme for industrial hard-coal combustion process in a power plant, accurate determination of an unmanned underwater vehicle's hydrodynamic parameters, self-sufficiency (autonomy) of space vehicle flight in conditions of supernumerary situations, for diagnostics of petroleum, for determination of a velocity profile of a fluid flow in cross-section of the pipe line with the help of a ultrasonic method, for segmentation of the complex images. About smart sensors and MEMS, see, please on http://www.sensorsportal.com/

5 Conclusions

Conclusions

? What you must believe in!

ANN technologies could be considered as a powerful tool, which supports high contact of experiment, theory, computer simulation and calculation.

6 Introduction

Introduction

What we do not know else?

How do ANN work in socio economical areas.

7 The goal

The goal

The goal of the seminar is: Share an experience Discuss recent and future ANN application in socio economic areas And of course - Learn how ANN can be applied to your data (if you have it)

8 Introduction (main

Introduction (main

No Life without Neural Networks !

9 Data Base

Data Base

You know there are three ways for creation Data Base: Real experiment Numeral experiment Analytical Solution Item 2 and 3 may not be used directly because math equations do not reflect a real life, because there are too many variables that reflect a real life.

10 Data Analysis

Data Analysis

What is the best way for analysis of real experimental data you know! It is Artificial Intelligence (Knowledge Discovery Data Base, Data Mining, Artificial Neural Networks)

11 Ex-1: Marriage and Dissolving of Marriage

Ex-1: Marriage and Dissolving of Marriage

12 Ex-1: Marriage and Dissolving of Marriage

Ex-1: Marriage and Dissolving of Marriage

13 Ex-1

Ex-1

14 Ex-1

Ex-1

15 Ex-1

Ex-1

16 Ex-1

Ex-1

17 Ex-1

Ex-1

18 Ex-2: BasegroupLab Development

Ex-2: BasegroupLab Development

Credit scoring solutions (risk assessment) Analysis of Moscow real estate market. The goal is to estimate the investment attractiveness of the projects

19 Ex - 2

Ex - 2

Forecasting the Financial Markets Market segmenting methods Market behavior modeling Analysis of the Ryazan city mass-media market with the help of Kohonen self-organizing maps

20 Future

Future

Educational System

Diagnostics of a modern state, analysis, and forecasting of dynamics of a system of education (on a level of an University, and a level of a region, and of the level of Russia. From the point of view of ANN, a main condition is a good Data Base. It is a very much difficult task requiring of usage of the advanced technologies of collecting (Data Warehouse). But even this technology can not ensure the solution of a set of problems

21 Future

Future

Ed-System - 1

Alongside with the DW, it is necessary to use and other ideas. Here, from the point of view of learning of separate sub processes in an Ed-System, usage of mathematical models of different mechanical, physical and chemical phenomena is perspective

22 Future

Future

Ed-System - 2

For example, an interesting model of self-organizing of a system during a transition of slow combustion to detonation exists in a combustion science. This model consists of both positive and negative feedback, both a recovering and destruction of a system. It can be utilized for creation of invented Data Base for modeling of self-organizing (self-recovering and self-destruction) of a system of education

23 Future

Future

Terrorism Modeling

Terrorism prevention problems: discovering of dangerous situations, forecasting of them as well as their management are in the center of world community attention for all XX century and XXI also The resolution of the General Assembly of the United Nations and the Resolution of Security Council 1368 and 1373 dated September, 12 and 28, 2001 called all the countries - members of the United Nations to cooperation in struggle against the international terrorism There is the special NATO Science Program [1] dealing with these problems. But we can mark here that many scientific meetings dealing with these problems are focused on particular concrete problems and strategies rather than general problems and strategies related to terrorism

24 Future

Future

Terrorism Modeling

But we have no solutions till now What is the reasons? May be the methods we have used before for modeling are insufficient? Terrorism is a complex socio economic system (SES). Terrorism is a complex socio economic phenomenon (SEP) involving political, religious, socio, economic, cultural as well as others aspects Therefore a solving of terrorism prevention problems demands a development of new methods of data revealing and collection as well as modeling of terrorism by means of most modern methods (KDD, DM, ANN) as well as Artificial Intelligence (AI) as a whole

25 Future

Future

Terrorism Modeling

Among concrete tasks we can mark here as the first the data collection (Data Warehouse) through literature sources (news articles, Internet materials, technical analyses, expert experience in particular from a point of a creation of intended scenarios of terroristic attacks, etc.). During that, simultaneously with that, we need to develop new AI terrorism models, explicate consistent, logical, and rigorous relationships between all aspects of terrorism, reveal roots of terrorism, its contemporary trends, create models of terrorist organizations, uncover early indicators of future terrorist actions, etc

26 Future

Future

Terrorism Modeling

The construction of terrorism models is very hard work and an enough long way. But, unfortunately, as it would think, we cannot hope to win the terrorism in a nearest future. Therefore, in the long run, real success in the war on terror requires building of real working terrorism models.

27 Conclusions

Conclusions

R&D Projects - 1

The creation of a global ANN model for automatic and fast (in real-time) solution of “Inverse and direct problems of optics” for all integral optical and non-optical methods of diagnostics, testing, and control by means of one-point measurement Application area is everywhere were optical and non-optical techniques is using Devices will become simpler and faster

28 R&D Projects - 1

R&D Projects - 1

We need. To create a `Data Base of input and output values of “Inverse and direct problems of optics” by means of solution of direct task, based on simple numeral calculation of integral equations. That have to be done for all possible symmetry of objects. After that we will be able to create the global ANN model for automatic and fast (in real-time) diagnostics, testing, and control by means of one-point measurement of signal to be registered for various optical and non-optical methods. Addition. In the framework we can solve a task of remote sensing by means of laser scattering technique (for example for ecological monitoring of environment), because the approach to be used is the same

29 R&D Projects - 2

R&D Projects - 2

The development of an automatic system for features extraction and recognition oscillograms (patterns, waveforms) for automatic and fast (in real-time) diagnostics, testing, and control of engineering devices, for smart sensors, soft sensors, smart watch by means of ANN (Data Mining tools)

30 R&D Projects - 2

R&D Projects - 2

We need. To create a Data Base of a set of input parameters (set of examples) that correspond to various state of a device to be diagnosed, tested and controlled. That has to be done for an every kinds of device during its working. After that we will be able to create the ANN model for automatic and fast (in real-time) diagnostics, testing, and control by means of well-known procedures of ANN training. After that we will be able to make an ANN based chip to diagnostics, testing and control. Addition. In a case of PC-based ANN we will be able to make an ANN based tool to diagnostics, testing and control that can be adapt during life cycle of device.

31 The first results of R&D - 2

The first results of R&D - 2

32 Other R&D Projects

Other R&D Projects

The NUS as a exclusive Asia representative of BasegroupLab (students, postgraduate students, and professionals teaching as well as promotion) in the areas of science and engineering, banking and financial activity, market and actuary

33 Other R&D Projects

Other R&D Projects

Development of DDT control system for PDE as well as for other devices (engines) and areas where special kind of electrical discharges (with a stabilization upon current or voltage) could be used

34 Other R&D Projects

Other R&D Projects

There are a very large Data Base of propellant combustion experimental data (in Russia). There are other Data Base like above (USA, Japan) But it have no Artificial Intelligence (Data Fusor, Data Mining, ANN, etc) If we could do that the price of the Data Base grows on 2 orders! The scientific values on 3 orders!

35 Other R&D Projects

Other R&D Projects

Development of nano size devices and coverings based on line-chain carbon for microelectronic (Field Emission Displays, X-ray tubes, Light Emitting Devises, etc) as well as Medicine

36 Now your practice

Now your practice

Do you have own data?

37 Conclusion

Conclusion

Thank you very much for your attention! See you later! P.S. Do you like play? Play with your Artificial Neural Networks Life Device. You have no it? Create! It help you to live! Have a nice day!

«The future perspectives of applications of artificial neural networks for diagnostics, modeling, and control in science and industry (methods, results, demonstrations hands-on)»
http://900igr.net/prezentacija/anglijskij-jazyk/the-future-perspectives-of-applications-of-artificial-neural-networks-for-diagnostics-modeling-and-control-in-science-and-industry-methods-results-demonstrations-hands-on-70219.html
cсылка на страницу

Курсы английского

25 презентаций о курсах английского
Урок

Английский язык

29 тем
Слайды
900igr.net > Презентации по английскому языку > Курсы английского > The future perspectives of applications of artificial neural networks for diagnostics, modeling, and control in science and industry (methods, results, demonstrations hands-on)