Available via license: CC BY 3.0
Content may be subject to copyright.
IOP Conference Series: Materials Science and Engineering
PAPER • OPEN ACCESS
Block Chain Financial Transaction Using Artificial Neural Network Deep
Learning
To cite this article: Bowen Zhao et al 2020 IOP Conf. Ser.: Mater. Sci. Eng. 740 012161
View the article online for updates and enhancements.
This content was downloaded from IP address 213.182.195.110 on 17/03/2020 at 17:15
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd
MEMA 2019
IOP Conf. Series: Materials Science and Engineering 740 (2020) 012161
IOP Publishing
doi:10.1088/1757-899X/740/1/012161
1
Block Chain Financial Transaction Using Artificial Neural
Network Deep Learning
Bowen Zhao1,*, Runyu Wang2, Yan Cai3, Enyu Zhao4
1School of Computer Science and Engineering, University of New South Wales,
Sydney, 2052, Australia
2Antai College of Economics & Management, Shanghai Jiao Tong University,
Shanghai, 200000, China
3School of Economics, Fudan University, Shanghai, 200433, China
4School of Economics, Renmin university of China, Beijing, 100872, China
*Corresponding author: Bowen Zhao, e-mail: bowen.zhao@unswalumni.com
Abstract. In order to adapt to the rapid development of the trading market in the financial
industry under the condition of artificial intelligence and solve the unstable problem of capital
loss in the process of financial transactions, in this study, firstly, the concept and principle of
block chain are summarized, and the problems that block chain can solve are analyzed in an all-
round way. Secondly, the concepts and computational principles of artificial neural network and
deep learning are analyzed, and they are integrated into block chains for artificial intelligence
payment. Finally, the block chain, which integrates artificial neural network and deep learning,
is applied to the financial market. The results show that block chains can make use of
decentralized networks to simulate virtual consumer payments, and can be used in many
financial fields. They can track and check information and prevent goods from being tampered
with in the logistics process. Artificial neural network (ANN) can imitate the human brain to
process and calculate the captured data by using neurons in both forward and reverse
transmission modes. Deep learning can process and recognize graphics and language information
by means of computer network. After updating the artificial neural network and the block chain
of deep learning, all kinds of transaction needs can be met through appropriate calculation
methods, so as to accurately grasp the trading situation of financial market and provide the
standard data of artificial intelligence to customers. This shows that the block chain which
integrates artificial neural network and deep learning plays a decisive role in financial
transactions, can bring huge potential benefits to the financial industry, and make the financial
transaction market develop steadily under huge competitive pressure.
1. Introduction
With the rapid development of science and technology, computer networks have been used in various
industries. Block chain is a newly emerging network technology, which has the characteristics of
common recognition mechanism and secret computing [1], and has been widely used in various
industries. Since the reform and opening up, China has always emphasized the rejuvenation of the
country by science and technology, and in 2015 launched a plan to use the Internet to reform traditional
industries [2]. After the Davos Forum in 2016, China has already implemented the economic
MEMA 2019
IOP Conf. Series: Materials Science and Engineering 740 (2020) 012161
IOP Publishing
doi:10.1088/1757-899X/740/1/012161
2
transformation plan and made clear the consensus of using science and technology to promote economic
development.
After years of development, artificial intelligence has been widely used all over the world, and
artificial neural network and deep learning network technology [3] as the frontier technology of science
has been used in various fields. Artificial neural networks and deep learning are also indispensable to
the development of new financial transactions, and they have attracted great attention all over the world.
By 2016, there are more than 7,000 financial trading enterprises based on deep learning, including block
chains, financing transactions, large data analysis and network data security modules [4]. Current
financial transactions have changed from passive resistance to active innovation.
In this study, the principle and application scope of block chain technology at the present stage are
elaborated, and the concepts, principles and application fields of artificial neural network and deep
learning are analyzed. The application of artificial neural network and deep learning to block chain
technology in financial transactions provides a creative method for the development of the financial
industry, which can enhance the competitiveness and core value of the financial industry, and has
practical feasibility.
2. Block chain technology
2.1. Concept and principle of block chain technology
Firstly, the concept of block chain technology refers to the consumption payment model which uses
decentralized network to simulate virtual consumption.
Decentralized network [5] refers to the integration of each scattered point in the same network, the
sharing of data and information, and then the related calculation and statistics for the network. Each
scattering point is relatively fair. It can input data information to the network and obtain data information
transmitted from other scattering points at the same time. Therefore, decentralized networks are also
called "mutual justice networks" [6]. The optimal feature of decentralized networks is that even if any
decentralized point is destroyed, the integrity of the network will be maintained and the network will
work normally without any influence. Only when all the scattered points of the network are destroyed
can it be in danger. The decentralized network structure is shown in the following figure:
Decentralized database Decentralized database
Decentralized database Decentralized database
Decentralized database Decentralized database
Figure 1. Decentralized network architecture
Secondly, the design of block chain technology is based on the fact that there are many information
modules with time stamp in the chain. Each information module has time difference between before and
after, and is constrained by Hash value [7]. Bitcoin is used as an incentive in the labor process, using
algorithms given by customers to prove and describe payment to shared accounts. Bitcoin [8] realizes
the conversion of payment function and no longer requires secondary payment. Before the block-chain
MEMA 2019
IOP Conf. Series: Materials Science and Engineering 740 (2020) 012161
IOP Publishing
doi:10.1088/1757-899X/740/1/012161
3
payment function appears, digital capital circulation needs supervision and management, so there must
be a third supervision unit, namely secondary payment [9]. However, nowadays, block chain payment
can share resources face to face, and the degree of confidentiality is very high. The combination of
public key and personal key greatly enhances the security of block chain payment. The two-way
encrypted fund holders are recorded in the block chain record book, which must be kept confidential
and the users who have reached a common agreement can pass through. Each scattered point is
connected through a special network, so users can check in the block browser.
Block chain is essentially a task checking reaction chain, that is, a sequential and indestructible
accounting book. Without the central characteristics, block chains can be used not only for financial
transactions, but also for recording all the content that customers need to record. The only requirement
is that the contents to be recorded can be represented by symbols. The contents of these records are
transaction records from one place to another, including transaction vouchers, which are appended to
the notebook. Next, these credentials are evaluated by block chains. Only credentials included in the
identified block links can be stored and cannot be broken or changed. The working steps of the block
chain are as follows:
New Data Records
Broadcast All-Network Verification of data Achieve consensus on block
correctness Storage of data
Figure 2. Block chain working steps
The workflow between the scattered points in the block chain is as follows. Firstly, when financial
payment is generated, the content of the payment will be distributed to each network. Each decentralized
point stores and integrates the accepted payment content. Secondly, each dispersed point can analyze
the data at the same time and get the opportunity to build block chains. After the establishment of
qualified block chains, the collected data are dispersed throughout the network. Finally, each block chain
is connected at the end to obtain a complete block chain. In this whole process, there is no need to verify
the steps, nor need third-party supervision. The only need is a complete reputation network and
consensus.
2.2. The application field of block chain technology
At present, block chain technology is highly valued in various fields. Beijing, Shanghai, Shenzhen and
other regions [10] include block chain technology in the scope of financial transactions, and highly
integrate it with the offline economy. China also includes it in the 13th Five-Year Plan in 2016. The
application fields of block chains in China are shown in the following table:
Table 1. Application field of block chain technology in China
Practical application
Bank
Combination of credit transaction and block chain
technology
CITIC
Combination of credit reporting and block chain technology
Ping an insurance
Combination of foreign trade and block chain technology
China merchants bank
Combination of account arrangement and block chain
technology
Weizhong
At present, the application of block chain technology in China includes many industries. There are
more than fifty A-share listed industries, and more than half of them are in the network field, which
means that block chain technology itself has a deep connection with computer networks. Block chain
technology is listed in A-share companies as follows:
MEMA 2019
IOP Conf. Series: Materials Science and Engineering 740 (2020) 012161
IOP Publishing
doi:10.1088/1757-899X/740/1/012161
4
Signal communication Computer Electronics traffic electric appliance media Light manufacturing
0
5
10
15
20
25
Number of enterprises
A-share Enterprise Combination Block Chain
Figure 3. Number of A-share listed companies combined with blockchain
2.3. Problems solved by block chain in financial transactions
The first is the innovation in tracking and checking information.
Through the integration of regional chain technology, all data in handover and payment can be more
effectively and publicly entered. At the same time, under the supervision of common participants, the
security and stability of data can be ensured, thus ensuring the integrity of logistics data.
The second is to prevent goods from being tampered with in the process of logistics.
Block chain technology is integrated into the whole flow process of cargo data, which can understand
the output and input information of each link in the whole process in real time, and can effectively
prevent cargo tampering on the basis of grasping the actual data. For example, in the process of mobile
phone transportation, all the cargo code circulation data should be mastered and recorded, in order to
prevent problems in the cargo need to recall, so as to be more flexible. Block chain technology can be
used to grasp logistics information at any time to meet the various requirements of customers.
3. Application of artificial neural network and deep learning in block chain
3.1. Artificial neural network
First, the human brain can analyze the captured data and store the information it needs in the brain.
Artificial Neural Network (ANN) is a computational model, which uses the information collected in the
network as a basis and imitates the human brain for data processing. ANN consists of innumerable
neurons. Its import layer receives data and then transmits it to the next level. The data is analyzed and
calculated step by step. In the 1990s, scientists added BP algorithm to the artificial neural network, so
that the problem of computing mode appeared in the calculation process of the network can be solved
smoothly. After that, the ANN develops rapidly. Nowadays, ANN has created many mesh models, the
most representative of which is BP neural network.
Secondly, the basic unit of ANN is neuron, which has many import windows, but only one export
window. After a series of training and data analysis, the information is finally calculated to 0 or 1 of the
exported data, so that the information can be summarized. The neurons are shown in the following figure:
MEMA 2019
IOP Conf. Series: Materials Science and Engineering 740 (2020) 012161
IOP Publishing
doi:10.1088/1757-899X/740/1/012161
5
N1
N2
N3
a
m
m
m
b
Figure 4. Neuron structure chart
Among them, N1, N2 and N3 represent imported values, m represents weights, a represent paranoid
and computational functions, and b represent derived values.
Finally, the calculation of data between neurons is divided into forward transmission calculation and
reverse transmission calculation, and finally the derived value is obtained.
3.2. Deep learning
Computer network is widely used in deep learning. Its tasks can be divided into graph recognition,
purpose verification, language meaning analysis and so on.
Firstly, graphics recognition is mainly based on the meaning of graphics response to identify graphics,
generally there is a series of stable coding. It is difficult to derive the encoding that corresponds to this
figure by the way of calculation in computer. After years of research, scientists have successfully solved
the problems encountered in graph classification in the field of deep learning and computing.
Secondly, the purpose of verification is to find the specified types of voice content in digital graphics
and videos. The process covers both purpose identification and classification. The most representative
method of purpose verification is RNCC method, which marks many small blocks on the alternative
graph. Next, these blocks are marked and summarized, and finally the orientation of the figure that needs
to be distinguished is obtained.
Finally, the analysis of linguistic meaning is to change the original imported graphics into
representative blocks, and classify all the materials into different categories and mark them, that is, to
identify the same type of blocks on the premise of linguistic meaning analysis.
3.3. Application of block chain block chain combining artificial neural network and deep learning in
financial transaction
With the development of financial industry and science and technology industry, the whole world is
paying attention to the integration of academic achievements and actual financial transactions. Block
chains and artificial intelligence have become the focus of current research. Artificial intelligence
includes the use of artificial neural networks and deep learning.
Firstly, the block chains of artificial neural networks and deep learning are updated by choosing
appropriate calculation methods to meet the needs of each transaction. In this study, the basic
configuration of the computer is optimized by using conventional computing, deep learning to identify
blocks, and self-updated neural network computing methods. Finally, the information bank is enriched
by self-input and self-update.
The method of calculation and fusion under the condition of deep learning calculation and monitoring
is used to remove some noise caused by monitoring, and finally a quantization network W with large
influence for deep learning parameters is formed. The internal learning criteria are W1 and W2,
respectively.
As far as the W1 learning rules are concerned, if the data imported and exported are:
1 1 2 2 N
m , , ,n
N
n m n m
(1)
The above equation can be expressed as follows:
MEMA 2019
IOP Conf. Series: Materials Science and Engineering 740 (2020) 012161
IOP Publishing
doi:10.1088/1757-899X/740/1/012161
6
( ) ( ) ( ) ( )
( )
1,1 1,1 1,1
11q QV l q QV l a l q QV l
= − − − −
(2)
In the equation, QV1,1 denotes the weight matrix, l denotes the input data, and q denotes the number
of weight matrices.
W2 learning rules are used after W1 to ensure the accuracy of calculation results. The calculation
method is as follows:
1
min , 1
qg
gq
ee
ee
−
+
(3)
eq and eg represent the Euclidean measurements of f*QV1,1 and g*QV1,1 in equation 2, respectively.
The equation for adjusting the weight matrix is as follows:
( ) ( ) ( ) ( )
( )
( )
1,1 1,1 1,1 1,1
11q QV l q QV l a l q QV l g QV l
= − − − −
(4)
When W1 and W2 are similar to the imported data, one is listed in the wrong data column, and the
other is listed in the correct column. At this time, the imported data will also be included in the function
formula, and the data next to it will be changed accordingly. In this way, W1 will make the weight of
the first analysis more precise, so that the calculated data will be more stable and accurate.
In conclusion, the principle of using artificial neural network and deep learning block chain
computing model in financial transactions is horizontal reverse transmission method.
Finally, the ANN and deep learning block chain model are applied to financial transactions. For
example, in the insurance industry, investment industry, etc., after the input of a pre-established
calculation model, artificial nerve simulation calculation begins. Each error parameter can be identified
by the model. Then noise elimination is adopted to assimilate and improve the accuracy of the
differences between block chains. In view of the differences among the blocks in the network, cluster
analysis can be used to verify the differences and connections between the blocks, and ultimately
complete the accurate calculation. For the insurance industry, the trading situation of the financial
market can be accurately grasped and the standard data of artificial intelligence can be provided to
customers.
4. Conclusion
Block chain is a skill and a creative service. Block chains based on ANN and deep learning can not only
provide intelligent contracts for financial transactions, intelligent management of property and funds,
but also enhance the transparency of financial funds circulation, making financial transactions stable
and safe.
In this study, the use of block chains in financial transactions is proposed, including tracking and
checking information and preventing goods from being tampered with in the logistics process. At the
same time, a block chain based on artificial neural network and deep learning is proposed, which can
improve the stability and security of capital flow in financial transactions, and enable the financial
industry to improve its own value and core competitiveness. The block chain, which integrates ANN
and deep learning, has strong practicability in the future trading process of the financial industry.
References
[1] Yao Q H, Xiao X. White Paper on the Development of Insurance Science and Technology in
China. China Insurance Science and Technology Laboratory, Fudan University, 2017, (14),
pp. 87-99.
[2] Xiao X. Block Chain and Insurance Innovation: Mechanisms, Prospects and Challenges.
Insurance Research, 2017, (05), pp. 43-52.
[3] Yao X D, Ke F. Innovation and Limitation of Block Chain Technology Application Sexual
Thinking. Rural Financial Research, 2016, (12), pp. 12-17.
[4] Yong Y, Wang F Y. Development Status and Prospect of Block Chain Technology. Kinetic
Chemistry, 2016, (4), pp. 482-493.
[5] Jian Z. Block Chain: Defining the New Financial and Economic Pattern in the Future. Beijing:
MEMA 2019
IOP Conf. Series: Materials Science and Engineering 740 (2020) 012161
IOP Publishing
doi:10.1088/1757-899X/740/1/012161
7
Machinery Industry Press, 2016, (08), pp. 15-28.
[6] Fu K L, Hui L, Xia Z. BP neural network prediction model and its application. Information and
Computer (theoretical version), 2016, (11), pp. 47-48.
[7] Si W. Design and Application of Discount Rate in Enterprise Value Assessment Based on BP
Neural Network.University, 2015, (09), pp. 110-120.
[8] HeD, Chen H M, Li Z S. Digital Recognition Based on Convolutional Neural Network. Journal
of Guizhou Normal University (Nature)Science Edition, 2017, (05), pp. 100-105.
[9] Wang Y S, Yao H X, Sun X S, Xu Pengfei and Zhao Sicheng. Study on the expressive ability of
self-encoder in deep learning. Computer Science, 2015, 42 (09), pp. 56-65.
[10] Qian L. Image Recognition Based on Deep Belief Network and Its Application. 2016, (09), pp.
101-109.