You are on page 1of 26

Privacy Preserving Using Back Propagation Neural Network For Cloud Computing

CHAPTER 1
1. INTRODUCTION
BACK-PROPAGATION is an effective method for learning neural networks and has
been widely used in various applications. The accuracy of the learning result, despite other facts,
is highly affected by the volume of high-quality data used for learning. As compared to learning
with only local data set, collaborative learning improves the learning accuracy by incorporating
more data sets into the learning process; the participating parties carry out learning not only on
their own data sets, but also on others’ data sets. With the recent remarkable growth of new
computing infrastructures such as cloud computing, it has been more convenient than ever for
users across the Internet, who may not even know each other, to conduct joint/collaborative
learning through the shared infrastructure.
Despite the potential benefits, one crucial issue pertaining to the Internet-wide
collaborative neural network learning is the protection of data privacy for each participant. In
particular, the participants from different trust domains may not want to disclose their private
data sets, which may contain privacy or proprietary information, to anybody else. In applications
such as healthcare, disclosure of sensitive data, for example, protected health information (PHI),
is not only a privacy issue but of legal concerns according to the privacy rules such as Health
Insurance Probability and Accountability Act (HIPAA). To embrace the Internet-wide
collaborative learning, it is imperative to provide a solution that allows the participants, who lack
mutual trust, to conduct neural network learning jointly without disclosing their respective
private data sets. Preferably, the solution shall be efficient and scalable enough to support an
arbitrary number of participants, each possessing arbitrarily partitioned data sets.
Challenges: Theoretically, secure multiparty computation (SMC) can be used to solve
problems of this kind. But the extremely high computation and communication complexity of
SMC, due to the circuit size, usually makes it far from practical even in the two-party case. To
provide practical solutions for privacy preserving Back-Propagation neural (BPN) network
learning, three main challenges need to be met simultaneously:
1) To protect each participant’s private data set and intermediate results generated during the
BPN network learning process, it requires secure computation of various operations, for

ISB&M School of Technology, Pune

Page 1

Privacy Preserving Using Back Propagation Neural Network For Cloud Computing

example, addition, scalar product, and the nonlinear sigmoid function, which are needed by the
BPN network algorithm.
2) To ensure the practicality of the proposed solution, the computation/communication cost
introduced to each participant shall be affordable. To accommodate a large range of collaborative
learning, the proposed solution shall consider system scalability. In particular, it shall be able to
support

an

arbitrary

number

of

participants

without

introducing

tremendous

computation/communication costs to each participant.
3) For collaborative training, the training data sets may be owned by different parties and
partitioned in arbitrary ways rather than a single way of partition.

ISB&M School of Technology, Pune

Page 2

For collaborative learning. neither the input data of each party nor the intermediate results (i. value of hidden layer node. we aim at enabling multiple parties to jointly conduct BPN network learning without revealing their private data.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing CHAPTER 2 2. During each learning stage.. weights. we consider a 3-layer (a-b-c configuration) neural network for simplicity but it can be easily extended to multilayer neural networks. PROBLEM DEFINITION In this paper. the main task for all the parties is to jointly execute the operations defined in the Feed Forward stage and the Back-Propagation stage. Pune Page 3 .e. except for the final learned network. The computational and communicational costs on each party shall be practically efficient and the system shall be scalable. Specifically. value of output layer node) generated can be revealed to anybody other than TA. ISB&M School of Technology. The input data sets owned by the parties can be arbitrarily partitioned.

al. This scheme provides strong protection for data sets including intermediate results. However. But the solution is proposed only for horizontal partitioned data. However. Pune Page 4 . LITERATURE SURVEY 3. Bansal et. which may also contain sensitive data. it just supports vertically partitioned data. enhanced this scheme and proposed a solution for arbitrarily partitioned data. just considers the two-party scenario though it supports arbitrarily partitioned data set. There still lacks an efficient and scalable solution that supports collaborative BPN network learning with privacy preservation in the multiparty setting and allows arbitrarily partitioned data sets. Moreover. just like. during the learning process. Chen and Zhong propose a privacy preserving BPN network learning algorithm for two-party scenarios. In practical implementation. ISB&M School of Technology. was proposed for the two-party scenario.1 Limitations Directly extending them to the multiparty setting will introduce a computation/communication complexity quadratic in the number of participants.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing CHAPTER 3 3. this enhanced scheme. To our best knowledge. none of existing schemes have solved all these challenges at the same time. such a complexity represents a tremendous cost on each party considering the already expensive operations on the underlying groups such as elliptic curves. RELATED WORK Several privacy preserving BPN network learning schemes have been proposed recently. Schlitter introduces a privacy preserving BPN network learning scheme that enables two or more parties to jointly perform BPN network learning without disclosing their respective private data sets.1.1. this scheme cannot protect the intermediate results. To overcome this limitation. Nevertheless. 3.

each possessing arbitrarily partitioned data sets. the participants jointly decrypt the results with which they update their respective weights for the BPN network. cloud servers learn no privacy data of a participant even if they collude with all the rest participants. we address this open problem by incorporating the computing power of the cloud. Pune Page 5 .1 INTRODUCTION The main objective of this project is to embrace the Internet-wide collaborative learning. and scalar product. The main idea of our scheme can be summarized as follows: each participant first encrypts her/his private data with the system public key and then uploads the ciphertexts to the cloud. As decryption of is limited to small numbers. it is imperative to provide a solution that allows the participants. we adopt the Boneh. our scheme makes the computation and communication complexity on each participant independent to the number of participants and is. During this process. the solution willl be efficient and scalable enough to support an arbitrary number of participants. we decompose most of the subalgorithms of BPN network into simple operations such as addition. Project Scope In this work. Goh. we introduce a novel design in our scheme such that arbitrarily large numbers can be efficiently decrypted. To support these operations over ciphertexts. To protect the intermediate data ISB&M School of Technology. For privacy preservation. SOFTWARE REQUIREMENT SPECIFICATION 4. who lack mutual trust. to conduct neural network learning jointly without disclosing their respective private data sets. highly scalable. thus. and Nissim (BGN) “doubly homomorphic” encryption algorithm and tailor it to split the decryption capability among multiple participants for collusion-resistance decryption. multiplication. Through offloading the computation tasks to the resource-abundant cloud.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing CHAPTER 4 4. cloud servers then execute most of the operations pertaining to the learning process over the ciphertexts and return the encrypted results to the participants. Preferably.

3 EXTERNAL INTERFACE REQUIREMENT There are many types of interfaces such as User Interface. Software Interfaces The system utilizes JDK framework which provides it with the necessary components to build system components and objects. 4. we introduce a novel random sharing algorithm to randomly split the data without decrypting the actual value. So in our proposed system we have design and developed GUI to interact with the system by using Java Swing classes. Thorough security analysis shows that our proposed scheme is secure. They can retrieve the same without letting the cloud server know about the data stored . 4.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing during the learning process. Software Interface and Hardware Interface. User Interfaces The user interface for the software shall be compatible to windows operating system. ISB&M School of Technology. They can ask for another party's data. Pune Page 6 . Communication Interfaces Graphical user interface is most convenience way to do the interaction with the system. plus providing the system with the required data access components.2 SYSTEM FEATURES General users This type of user are the clients. Cloud Sever Cloud server is a server where the data of various parties will be stored .

disclosure.4 FUNCTIONAL REQUIREMENT            Business Rules Transaction corrections. Extensibility  Extensibility allows adding new component to the system. disruption. Better component design: To get better performance at peak time Security Requirements  Secure access of confidential data (user’s details). replaces the existing ones. adjustments and cancellations Administrative functions Authentication Authorization levels Audit Tracking External Interfaces Certification Requirements Reporting Requirements Historical Data Legal or Regulatory Requirements 4. Pune Page 7 .Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 4.5 NON FUNCTIONAL REQUIREMENT Performance Requirements  High Speed: System should process voice messages in parallel for various users to give  quick response then system must wait for process completion. Flexible service based architecture will be highly desirable for future extension ISB&M School of Technology. This is done without affecting the components that are in their original places. modification or destruction. Information security means protecting information and information systems from unauthorized access. use.

Availability. It refers to the ability of technical support personnel to install. and provide hardware or software maintenance in pursuit of solving a problem and restoring the product into service. Serviceability  In software engineering and hardware engineering.6 ANALYSIS MODEL 4. debug or isolate faults to root cause analysis. 4.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing Scalability  The solution should be able to accommodate high number of customers and brokers.1 DFD level 0 Figure 1 ISB&M School of Technology. Compatibility  Compatibility is the measure with which user can extend the one type of application with another. configure. is one of the aspects (from IBM's RASU (Reliability. and monitor computer products. Both may be geographically distributed. identify exceptions or faults. Pune Page 8 . Serviceability.6. serviceability also known as supportability. and Usability).

Pune Page 9 .3 DFD level 2 Figure 3 ISB&M School of Technology.6.6.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 4.2 DFD level 1 Figure 2 4.

6.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 4. Pune Page 10 .4 Class Diagram Figure 4 ISB&M School of Technology.

Pune Page 11 .6.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 4.5 ER Diagram Figure 5 ISB&M School of Technology.

Objective And Planning 22nd August 2014 28th August 2014 4. The minimum configuration required for system is: 4. Literature Survey 8th August 2014 21stAugust 2014 3.7. Windows XP onwards. Pune Page 12 .Hardware Requirements & Budget 29th August 2014 5th September 2014 5. Algorithm Analysis 15th September 2014 21st October 2014 ISB&M School of Technology.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 4.2 Software  1 gigahertz (GHz) or faster 32-bit (x86) or 64-bit (x64) processor  1 gigabyte (GB) RAM (32-bit) or 2 GB RAM (64-bit)  16 GB available hard disk space (32-bit) or 20 GB (64-bit)  DirectX 9 graphics device with WDDM 1. Topic search and Finalization 4th July 2014 25th July 2014 2.1 and above : DATABASE Jdk 3 and above : For JAVA NetBeans 6 and above : project coding environment GlassFish 5. DFD . Planning Start Date Completion Date 1. No.7 SPECIFIC REQUIREMENTS Project System will be Windows based supporting versions.0 or higher driver 4. UML diagrams 8th September 2014 10th September 2014 6.8 SYSTEM IMPLEMENTATION PLAN Sr. Algorithm Design 10th September 2014 13thSeptember 2014 7.1 Hardware     MySQL 5.1 and above : server hosting 4.7. Software .

Working Model and Testing. Study of Project related Technology 1st November 2014 3rd January 2014 10. No. 16th March 2014 29th March 2014 13.9BUDGET Sr. Pune Page 13 . Preliminary Report 22nd October 2014 29thSeptember 2014 9. 24th February 2014 15th March 2014 12. Coding and Implementation of Project. Final Report and Deployment of Project. 2nd April 2014 20th April 2014 Table 1 4. NetBeansIDE 1 OSS 4. 5th January 2014 20th February 2014 11. Computer 1 22000 2. Product Quantity Cost 1 1. JDK 1 OSS Total 27000 Table 2 ISB&M School of Technology. Tested and Executable Project Model. Windows XP/Windows 7 1 3000 3.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 8.

SYSTEM DESIGN 5. 1. Figure 6 5. Cloud Server 2. BPNN is used by Clients to train their data.1 SYSTEM ARCHITECTURE: In our Proposed System We Have 2 Main Blocks. Only Clients can have knowledge of data stored on Cloud. Pune Page 14 .2 UML DIAGRAMS ISB&M School of Technology. Clients Cloud Server is used to store the data of Clients. The data is in encrypted form.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing CHATPER -5 5.

Use case diagrams are also important for testing executable systems through forward engineering and for comprehending executable systems through reverse engineering. subsystem. Use case diagrams are important for visualizing. or class.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 5.2. Each one shows a set of use cases and actors and their relationships. and collaboration diagrams are four other kinds of diagrams in the UML for modelling the dynamic aspects of systems).2. sequence diagrams.1 Use-Case Diagram Use case diagrams are one of the five diagrams in the UML for modelling the dynamic aspects of systems (activity diagrams. or modelling the requirements of the behaviour of these elements. Use case diagrams are central to modelling the behaviour of a system. state chart diagrams. subsystems. this involves modelling the context of a system. and documenting the behaviour of an element. or a class. a subsystem. specifying. Pune Page 15 . They make systems.2 Activity Diagram ISB&M School of Technology. For the most part. Figure 7 5. You apply use case diagrams to model the use case view of a system. and classes approachable and understandable by presenting an outside view of how those elements may be used in context.

Class diagrams that include active classesaddress the static process view of a system Figure 8 5. interfaces.3 Sequence Diagram ISB&M School of Technology. Classdiagrams address the static design view of a system. Pune Page 16 . and collaborations and their relationships.These diagrams are the most common diagram found in modeling object-oriented systems.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing A class diagram shows a set of classes.2.

which is especially useful in modelling reactive systems. They are especially important in modelling the behaviour of an interface. transitions. Figure 9 CHAPTER -6 ISB&M School of Technology. You use statechart diagrams to illustrate the dynamic view of a system. Statechart diagrams emphasize the event-ordered behaviour of an object. and activities. consisting of states. Pune Page 17 . or collaboration. events. class.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing A statechart diagram shows a state machine.

JavaServer Faces. RMI. servlets. This allows developers to create enterprise applications that are portable and scalable. dual-licensed under two free software licences: the Common Development and Distribution License (CDDL) and the GNU General Public License (GPL) with theclasspath exception.1 TECHNOLOGY USED IN PROJECT 6. with an added component called Grizzly which uses Java New I/O (NIO) for scalability and speed.2 GlassFish GlassFish is an open-source application server project started by Sun Microsystems for the Java EE platform and now sponsored byOracle Corporation.1. GlassFish runs straight on top of the Apache Felix implementation. Pune Page 18 . Optional components can also be installed for additional services. It code released uses a by derivative Sun and Oracle of Apache Tomcat as the servlet container for serving Web content. JMS. The supported version is called Oracle GlassFish Server.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 6. It also runs with Equinox OSGi or Knopflerfish OSGi runtimes. TECHNICAL SPECIFICATION 6. test and debug Java applications and applets. JPA. GlassFish is based on source Corporation's TopLink persistence system. HK2abstracts the OSGi module system to provide components. Built on a modular kernel powered by OSGi. GlassFish is the reference implementation of Java EE and as such supports Enterprise JavaBeans.JavaServer Pages. etc.1. Such services can be discovered and injected at runtime. 6. GlassFish is free software. which can also be viewed as services. and that integrate with legacy technologies. Features of GlassFish ISB&M School of Technology.1 JDK JDK (Java Development Kit) is a free software development package from Sun Microsystems that implements the basic set of tools needed to write.

 Allow decryption of arbitrary large messages.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing  A lightweight and extensible core based on OSGi Alliance standards  A web container  An easy-to-use Administration Console for configuration and management  Update Tool connectivity for updates and add-on components  Support for high availability clustering and load balancing 6. Pune Page 19 .  Collaborative BPN network learning over arbitrarily partitioned data. 6.  Cost independent of number of parties.3 APPLICATIONS      Our system can be used for It can also used to control computer It can be used to control various software applications Our system can be used to control power point presentation Our system can be used to control or play games CHAPTER-7 ISB&M School of Technology. efficient and secure.2 ADVANTAGE  Provides privacy preservation for multiparty.  Scalable.

In our proposed approach. CONCLUSION In this work.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing 7. One interesting future work is to enable multiparty collaborative learning without the help of TA. The cloud can execute most operations pertaining to the BPN network learning algorithm without knowing any private information. Pune Page 20 . This work tailors the BGN homomorphic encryption algorithm to support the multiparty scenario. Complexity and security analysis shows that our proposed scheme is scalable. the parties encrypt their arbitrarily partitioned data and upload the ciphertexts to the cloud. and secure. CHAPTER -8 ISB&M School of Technology. we proposed the first secure and practical multiparty BPN network learning scheme over arbitrarily partitioned data. efficient. The cost of each party in our scheme is independent to the number of parties. which can be used as an independent solution for other related applications.

20(10):1554-1564. Member. Trans. Neur. [3] PRIVACY PRESERVING BACK-PROPAGATION NEURAL NETWORK IN CLOUD COMPUTING International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 3 Issue 6..Privacy Preserving Using Back Propagation Neural Network For Cloud Computing REFERENCES [1] Privacy Preserving Back-Propagation Neural Network Learning Made Practical with Cloud Computing. Jiawei Yuan. Netw. Chen and S. U. Student Member. 2014 [2] Privacy Preserving Back-Propagation Neural Network Learning Over Arbitrarily Partitioned Data Ankur Bansal Tingting Chen Sheng Zhong Computer Science and Engineering Department State University of New york at Buffalo Amherst. Pune Page 21 . June 2014 [4] PRIVACY PRESERVING BACK-PROPAGATION NEURAL NETWORK LEARNING. Oct. NY 14260. A. and Shucheng Yu. T. IEEE. 2009 ANNEXURE ISB&M School of Technology. S. IEEE. Zhong.

Pune Page 22 // backward pass . ex) // forward pass actual = teacher-output(ex) compute error (prediction .actual) at the output units compute \Delta w_h for all weights from hidden layer to output layer // backward pass compute \Delta w_i for all weights from input layer to hidden layer continued ISB&M School of Technology.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing ANNEXURE A: Project Analysis of Algorithm Design Project Analysis: BPNN Figure 10 initialize network weights (often small random values) do forEach training example ex prediction = neural-net-output(network.

For our system. ANNEXURE B: Project Quality and Reliability Testing of Project Design ISB&M School of Technology. ___________________________________________________________________________ P Complete Problem A problem is in P class if it is solvable in polynomial time by a deterministic algorithm.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing update network weights // input layer not modified by error estimate until all examples classified correctly or another stopping criterion satisfied return the network COMPLEXITY The complexity of the back propagation neural network depends on number of input layers. algorithms are deterministic and the overall complexity is O(n) which shows that it is in P class. Thus for n layers. Pune Page 23 . the complexity is – ___________________________________________________________________________ O(n).

 Grey Box Testing: Grey Box testing is a technique to test the application with limited knowledge of the internal workings of an application.  White Box Testing: White box testing is the detailed investigation of internal logic and structure of the code. There are different methods which can be use for Software testing. errors or missing requirements in contrary to the actual desire or requirements. This activity results in the actual. White box testing is also called glass testing or open box testing.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing Testing is the process of evaluating a system or its component(s) with the intent to find that whether it satisfies the specified requirements or not. There are different types of testing which may be used to test a Software during SDLC.  Black Box Testing: The technique of testing without having any knowledge of the interior workings of the application is Black Box testing. is when the tester writes scripts and uses another software to test the software. and Nissim BPN : Back Propagation Neural ISB&M School of Technology. expected and difference between their results. Pune Page 24 . Goh. ANNEXURE C ABBREVIATIONS B BGN : Boneh.  Manual testing: This type includes the testing of the Software manually i. In simple words testing is executing a system in order to identify any gaps.e.  Automation testing: Automation testing which is also known as Test Automation. without using any automated tool or any script.

Pune Page 25 .Privacy Preserving Using Back Propagation Neural Network For Cloud Computing BPNN : Back Propagation Neural Network C CDDL : Common Development and Distribution License D DFD : Data Flow Diagram E ER Diagram : Entity Relationship Diagram G GB : Giga Byte GH : Giga Hertz GNU : General Public License GUI : Graphical user Interface H HIPAA : Health Insurance Probability and Accountability Act I IDE: Integrated Development Environment J JDK : JAVA Development Kit JMS : JAVA Message Service JPA : JAVA Persistence API N NIO : New I/O O OSGI : Open Service Gateway Initiative OSS : Open Source Software P PHI : Protected Health Information ISB&M School of Technology.

Availability.Privacy Preserving Using Back Propagation Neural Network For Cloud Computing R RASU: (Reliability. Serviceability. and Usability) RAM : Random Access Memory S SMC : Secure Multiparty Computation T TA : Trusted Authority U UML:Unified Modelling Language ISB&M School of Technology. Pune Page 26 .