You are on page 1of 25

Running Head Emerging Accounting Technologies - Machine Learning 0

EMERGING ACCOUNTING TECHNOLOGIES - MACHINE LEARNING

Course Name: Information Technology

Group 5: Machine Learning

Group Members:

1. Avreet Kaur
2. Karan Dilip Patel
3. Sadia Binte Mahmod
4. Sayan Nandy
5. Vinayak Indolia

Submission Date: 21st November 2020


Emerging Accounting Technologies - Machine Learning 1

Executive summary

This report has been prepared to showcase innumerable amount of benefits of Machine Learning
(ML), which is a new emerging technology of the present era. ML solves different kinds of
complex problems, which human beings face in their day-to-day life. Description of the
problems faced by humans and their cure is shown in this report, by examples such as driving a
car automatically, detecting a fraud, recommending things based on someone’s interests, and so
on. Different techniques of ML and how they enable a computer to operate, has also been
covered in this report. The advantages and disadvantages of ML has been discussed along with
the potential future uses in accounting profession. The case study has been described in the
report of Tesla which uses the ML effectively.
Emerging Accounting Technologies - Machine Learning 2

Table of Contents
Introduction.................................................................................................................................................3
Overview of Machine Learning...................................................................................................................3
History of Machine Learning...................................................................................................................4
Effectiveness of Machine learning and its operational context................................................................5
Machine learning in accounting profession.................................................................................................6
Machine Learning Use cases in Accounting Profession:.........................................................................8
Impact of Machine Learning’s on Accounting Profession:....................................................................10
Advantage and Disadvantage of Machine learning....................................................................................11
Advantages............................................................................................................................................11
Disadvantages........................................................................................................................................12
Seriousness of disadvantages:................................................................................................................14
Ways to overcome the disadvantages:...................................................................................................14
Potential future uses..................................................................................................................................14
Case study.................................................................................................................................................17
Conclusion.................................................................................................................................................20
Reference...................................................................................................................................................21
Emerging Accounting Technologies - Machine Learning 3

Introduction
Nowadays, in the modern world, new emerging technologies have come a long way in
helping people solve their problems at ease. Likewise, Machine Learning (ML) is an emerging
technology, which makes computers learn spontaneously without any programming. Using
algorithms, it can access data automatically and use them whenever required. The use of machine
in accounting profession will be discussed in the report along with the advantages and
disadvantages in accounting profession. The potential future uses of machine learning will be
analysed and the innovative ways of using ML in accounting profession will be included in the
report. The report will include the real-life example of Tesla which uses machine learning in
manufacturing and operations of the products and services. The ways company uses the machine
learning is analysed along with the benefits which the company receives with usage of machine
learning.

Overview of Machine Learning


Machine Learning has turned out to be an effective technology which solves a lot of
problems for people in various fields. In comparison with doing explicit programming, to carry
out a particular task, ML is at a far more advantageous position. It does the work of humans, that
too, in an automatic way (Cawi, La Rosa & Nehorai, 2019). For instance, when an email arrives,
humans need to check the folders themselves. Approaching through ML eases this job as the
email is automatically detected in whichever folder it is, without use of any if-else statements.
ML uses different algorithms to complete a particular task by using data and instructions
given to them by humans. Data that is provided to computers by humans can be of two types,
data which have predefined output and data which does not have predefined outcome. ML has
shown its effectiveness in the case of both, where data with the predefined result is known as
supervised learning, and the other one is known as unsupervised learning (Kodelja, 2019). ML
uses several techniques to detect things such as something besides an individual in a picture,
name of a car in a picture, and many others.
Emerging Accounting Technologies - Machine Learning 4

History of Machine Learning


Machine Learning has turned out to be one of the best prestigious innovations, which
makes a computer carry out tasks that only a human can do. A computer, with help of ML, can
detect several things like things present in an image, airport scanning systems, drive vehicles
automatically without humans sitting in the driving seat, and so on (Lee et al. 2019). It has
evolved after a lot of research and examining whether a computer can learn for real.
● The year 1950 marks beginning of ML creation, when Alan Turing designed a Turing
Test to check if a computer has real intelligence or not. Provided it has, studies regarding
ML have proceeded.
● Getting to know that computers can learn, the year 1952 brought up first computer
program by Arthur Samuel. This received a lot of success, including involvement of IBM
computers in modifying it.
● In order to implement a similar thought process like a human brain, Frank Rosenblatt
designed the first-ever perceptron, that is a neural network of computers, in 1957.
(Poolwan & Smanchat, 2018)
● The year 1967 was very crucial, as it implemented an algorithm for computers. It was
termed as the "nearest neighbor” algorithm, used for pattern recognition.
● Year 1979 marked the innovation of “Stanford Cart”, which was made by students of
Stanford University, for purpose of navigating objects in a room. Furthermore,
Explanation Based Learning (EBL) and NetTalk (learning to utter words like a baby)
were invented in 1981 and 1985, respectively.
● The 1990s were all about the gradual shift of learning about machines in a data-driven
approach, which eventually led to computers learning to analyse large chunks of data and
perform tasks.
● The 2000s was about developing different other algorithms that could make computers
more prone to carry out various tasks which a human brain can do.

Year Incident

1950 Turing Test

1952 First computer program

1957 Perceptron invention


Emerging Accounting Technologies - Machine Learning 5

1967 Nearest Neighbor algorithm

1979 Stanford Cart

1981 EBL

1985 NetTalk

1990s Machine Learning in data driven approach

2000s Developing more algorithms


Table 1: Timeline
(Source: Surender, 2020)

Effectiveness of Machine learning and its operational context


Machine Learning has gradually become one of the most effective technologies among
the new emerging technologies of this modern world. Machine Learning makes use of different
algorithms, which takes a set of data, processes it, and gives desired results. Ml enabled
computers to work similar to human brains, being able to detect even minute details which only a
human being can. ML is highly effective as it lowered down human interaction, making
computers automatically carry out tasks. For instance, ML makes a car move automatically
without anyone driving, by means of applying algorithms of modern cars, or recommending
necessary requirements to humans as per their interests. ML turns out to be effective in business
as well, as many IT industries use the Internet-of-Things (IOT). ML coupled with IoT can act as
key equipment in solving complex problems related to business. Unsupervised learning
algorithms can be used in courts as well, to let computers decide independently. ML helps
organisations know from beforehand, which customers to target.
Machine Learning majorly operates by two different techniques, supervised and
unsupervised algorithms (Özer, 2018). Before beginning its operation, an algorithm is selected
that is appropriate for that particular task, in which the data is inputted. Data that is inputted into
computers are matched with the selected algorithm, which is essential. In order to test its
working, new data is fed again and its results, including what it predicted are checked.
Predictions can be correct or wrong, according to which algorithms are either re-trained or taken
into consideration.
Emerging Accounting Technologies - Machine Learning 6

Two of the familiar programming languages that are used in implementing Machine
Learning algorithms, are ‘R’ and ‘Python’ programming languages. Depending on what tasks to
perform, different languages are selected in the beginning. Although many languages are
available, Python is chosen almost every time because of its diversified features. Works such as
Data Analysis and Data Mining are carried out by Python which is an important step of ML.

Figure 1: Two Different Techniques of Machine Learning


(Source: Cawi, La Rosa & Nehorai, 2019)
This above diagram shows the two primary techniques of operation of ML, namely
supervised and unsupervised, as discussed in the previous paragraphs. Supervised technique is
subdivided into Classification and Regression, detecting discrete and continuous responses
respectively. By contrast to supervised technique, unsupervised technique is further subdivided
into Clustering. Clustering, which is the most popular unsupervised ML technique, is for purpose
of searching for any hidden patterns in data.

Machine learning in accounting profession


Machine Learning provides the system the flexibility to automatically learn and ad-lib from
experience without being unequivocally coded. It allows the systems to predict and make
changes to the data as humans would. It empowers the computer to do Machine-based Learning,
previously done by humans, however within the accounting profession, where people cater to
rote tasks, Machine Learning can be replacement of human capabilities for the better (Bräuning,
et al., 2017).
Emerging Accounting Technologies - Machine Learning 7

Companies to contour their business operations are currently embracing and


incorporating new technologies, and one that is on the highest of their list is Accounting. The
outcome is such as Machine Learning is providing positive results like enhanced accuracy,
increased productivity, and reduced cost. Accountants can rely on Machine Learning to analyze
data and help them refine forecasting models. In layman’s term, a poor quality data can lead to
an error down the line, but Machine Learning can enhance the quality of data along with
accuracy without an accountant having to put in much effort. The technology will flag
inaccuracies and errors and it will also prevent the occurrence of new errors based on what it
learns, thus improving and speeding up the process.

According to several surveys, nearly 80-85% of the individuals believe that Machine
Learning ends up in competitive advantage which leads to the improvement of the productivity
of the corporate companies. Machine Learning is being employed by several Accounting
Corporations where it analyses a vast amount of data at high speed, which might not be feasible
for humans.

Machine Learning helps to lighten the burden of tedious accounting on firms and their
business clients so that the accountants can focus on other important things such as providing
customers with quality services, creating a better working environment for the employees,
innovating for the advancement of the industry. A continuous improvement would include
automation of repetitive tasks which can fairly be achieved by the Accountants through Machine
Learning.

Machine Learning might be useful for easily automating the process of Accounting such as
generating Income/Expenditure Reports, Accounts Receivables/Payables and Assessment of risk.
With the Advancement of technology like Scanning of Documents, Software to match Source
Documents, the professions necessitating the manual handling of documents have by now started
vanishing (Stancheva-Todorova, 2018).

For example, Machine Learning can link and match an Invoice to Purchase Order on its receipt
and charge it to an expense account and position it in a pool for release of payments, and once it
is reviewed by an employee, release them for payments.
Emerging Accounting Technologies - Machine Learning 8

Machine Learning Use cases in Accounting Profession:

1.) Process Automation: Application of Machine Learning, being one of the most common,
in Accounting Profession is Process Automation. The technology allows replacement of
automation of repetitive tasks, manual work, and increased efficient productivity.
Machine Learning would enable a corporate company to cost optimization, develop
customer experiences, and boost up provision of services. Automation of Machine
Learning in Accounting Profession can be used in the form of:
 Chatbots – Machine Learning will resolve common queries for users with
efficiency like when bills are due, the most recent account balances, and
standings on Accounts. For Example – Through Facebook Messenger platform,
AI-driven Chatbot used by Wells Fargo connects with the users and aid them
with the passwords and accounts related information.
 Paperwork Automation – Digitalization tracks that which file was accessed by
whom and when and that will increase the protection of information and files.
Throughout the process, auditors will have simple access to the digital files for
processing and they need not search for file cabinets for documents. This will
increase the accuracy and potency of audits which will make it possible for
auditors to Audit 100% of the Firm’s Financial Transactions rather than simply
Samples (Haq, et al., 2020). For Example – JP Morgan Chase launched a
Contract Intelligence (COiN) platform to process legitimate documents and
extracts vital and important data from them. Their data showed that the
commercial credit agreements would take upto 360,000 labor hours for manual
Emerging Accounting Technologies - Machine Learning 9

review of 12,000 agreements. While Machine Learning would evaluate the


similar figure of agreements within a few hours.

2.) Security: Security threats in Accounting Profession are expanding with the rise in variety
of transactions, consumers, and third-party integrations. However, Machine Learning
Algorithms are exceptional at detection of Frauds. For Example – This technology has
been very fruitful for Banks to observe thousands of transactions for each account in real
time. The algorithm will examine the cardholders every action and assess if an
endeavored activity is the trait of that cardholder. This would spot deceitful behavior with
high preciseness (Cho, et al., 2020).

 If the system identifies Questionable Account Behavior, additional information


can be requested from users for Authenticating the transaction or even prevent
the transaction altogether. Machine Learning Algorithms need very little or no
time to analyze a transaction. The speed facilitates not only to detect frauds after
it is committed but also helps to prevent it in real time (Bakarich and O’Brien,
2020).

 Machine Learning can help monitor finance wherein the Data Scientist can train
the system to detect many Micro-payments and flag money laundering
techniques.

 For Example – Some notable Corporates like Adyen, Payoneer, PayPal, Stripe
and Skrill invest largely in Machine Learning Security.

3.) Underwriting and Credit Scoring: Machine Learning Algorithms corresponds


seamlessly with the underwriting assignments which are prevalent in Finance and
Insurance. In real life scenario, a well-trained system will perform the underwriting and
Emerging Accounting Technologies - Machine Learning 10

credit-scoring activities which would help employees work much faster and more
efficiently. For Example – BBVA Bancomer has collaborated with an alternate credit-
scoring platform Destacame. The process would include Destacame to access bill
payments information from the company via open API’s and using the behavior of these
bill payments, produce a credit score for a customer applying for a credit and send these
results to the bank for the approval of credit (Richins, et al., 2017).

Impact of Machine Learning’s on Accounting Profession:


I. Auditing

 Reducing audit personnel.


 Internal control focus.
 Improvised detection of frauds.

II. Tax Compliance

 Reduced time to review material.


 Machine learning expertise.
 Planning and Designing opportunities for New Taxes.
 Value-addition enhanced facilities.

III. Advisory Services

 Decreased Data Processing time.


 Improved precision and lowered costs.
 Focus on service and added value.
Emerging Accounting Technologies - Machine Learning 11

Advantage and Disadvantage of Machine learning

Advantages
1. Continuous improvement and automation:

Machine learning is the technology that is capable to learn from the data provider in the
system to accountant professional. ML is an effective model for accuracy and efficiency for
making decisions that improve the system. An example of continuous improvement is Amazon,
Walmart, etc. which helps in collecting the huge volume of new data daily. A huge amount of
training data is available through machine learning with the accuracy for finding associated
products or recommendations. Machine learning is a powerful utility that automates the bulk
information and data for decision-making tasks. The machine learning program saves time for
developers to calculate the data and use the time for more productive use. For example, the
calculation of risk and beta of the company based on historical data that is bulk in quantity can
be done in minimum time using the machine learning program (Ngiam, & Khor, 2019). Machine
learning plays an important for changing the world with the automation and improvement in
techniques.

2. Trends and patterns identification:

The machine learning program can be used in classification and regression problems using
variously supervised, unsupervised, and reinforced learning algorithms conducted by
professional accountant. The machine learning program can use various trends and patterns with
a huge amount of data for analysis by accountant. For example, a Machine learning program is
used by Amazon based on the buying patterns and search trends of the customers and predicts
products for different patterns (Chicco, 2017). The factors involved in the machine learning
program that make it reliable. Data handling is a major factor in machine learning. Machine
learning is having the biggest role in trends and patterns identification and data handling.

3. Wide range of application:

The machine learning program is used in a wide range of different industries such as
education, defense, stock market, groceries, manufacturing, etc. Business organizations use past
data and analyze trends and patterns for generating profits, cut costs, automate, and predict the
Emerging Accounting Technologies - Machine Learning 12

future which is used by different accountant professionals. The machine learning program is used
in different areas like GPS tracking for traffic, the filter of email data such as important, spam,
the predication of data, correction, and check of the spell, etc. which is part of today's scenario.
The machine learning program is an important part of Artificial Intelligence which helps in
founding the latest trends and applications. A machine learning program is a useful tool to
provide the education in the future as various creative techniques for students to study. For
example, a school in China has used machine learning for improving student focus (Chen, Li,
Knibbs, Hamm, Cao, Li, & Guo, 2018). The advertisement sector uses the data analyzed by ML
for providing the preferences and a great way for improving the e-commerce with Machine
learning. The ML can also use for scientific and technical data in the field of medical, business,
banking that helps in create different opportunities. The ML also helps in interactions with
different customers and detection of issues more quickly, helps in lifting-up businesses.
Therefore, investment in ML technology is worthy for the different sectors.

Disadvantages
1. High error susceptibility

The output of using the technology is fully dependent on the input inserted in the system
means if incorrect data inserted in the machine learning program then the output will be
incorrect. The data that is put in the models for the evaluation and solution of the problem must
be clean and accurate. The automated processes using machine learning become useless if the
data is improper and incorrect. The data inserted in the system is a manual process and there is a
high possibility of data will be incorrect that may cause incorrect results or errors. The example
is in case of the incorrect filter is set by the customers then the recommended trends and products
may be incorrect and is not as per customer choice. The machine learning program is based on
algorithms that are required for accurate results (Yeom, Giacomelli, Fredrikson, & Jha, 2018).
The main problem that arises in the ML is the training as well as testing of data. In case when
data is huge, sometimes removing errors is becoming impossible in the ML. The errors that
occur in the data may cause headaches to users and are also time-consuming.
Emerging Accounting Technologies - Machine Learning 13

2. Data acquisition and time consuming:

The acquisition of data is the most painful point in the field of data science and machine
learning. The collecting data is also costly to collect in monetary and timely measures. The
surveys are used for collecting data that may be bogus and incorrect as a large volume is
collected. While using the machine learning model, many situations arise when there is an
imbalance in data which leads to inaccuracy and incorrect data processing. Therefore, the data
collection is the biggest disadvantage of Machine learning. The ML model is work on data and
the huge amount of data which is inconsistent create problem in the outcome. The data is
required to keep updated and consistent. For effective results, the data must be updated and new
(Haider, Baig, & Imran, 2020).

Time-consuming is also an important disadvantage of machine learning as data collecting is a


time-consuming process along with the process of removing errors found in testing and training
of data is also a time-consuming process. The high amount of data, time to learn from data, and
processing of data required effective time. The ML algorithms may take more time for
processing and removing the error as think by the organization. If the data is large and advanced
then the system will take more time and also consume the power of systems which becomes
hectic and required more allotted space (Leavy, 2018). Therefore, ML models are not beneficial
based on time consumption.

3. Algorithm Selection and costly to implement:

The ML model use as well as implement various algorithms for identifying the solution
which is a manual and tedious task. The manual processing for the selection of the Algorithm is a
disadvantage of the ML model (Jiang, & Nachum, 2020). The selection of an algorithm is a run
and test of data and a manual job which is very much time-consuming.

The implementation of the ML model is costly for the business organization as the server
and techniques are required high cost to incur. The ML process is costly from the collection of
data, selection of algorithm, removing the error and test and run of data (Veale, & Binns, 2017).
The cost may become a disadvantage for the business organization if the results of ML models
may be inaccurate for the business purpose.
Emerging Accounting Technologies - Machine Learning 14

Seriousness of disadvantages:
The disadvantages described above are serious for the accounting professional as in case of high
error in outcome the purpose of using the machine learning program is not useful for the
professional and the saving of time is not possible due to high time consumption in selection of
Algorithm, collection of data and resolving the errors arise in the ML models. The disadvantage
is serious as they are reasons of decrease or eliminate the benefits of ML model for the
professional.

Ways to overcome the disadvantages:


The ways in which the professional can overcome the disadvantages in ML models:

● The professional should collect data from reliable sources that reduce the issue of
incorrect data and information.
● The selection of algorithm program is done based on minimum data and tests which
reduce the time for selection of algorithm.
● The data should be tested and training is done on lower amount of data for identify the
error in the information and program and after resolving the same update the complete
data in ML model. The same reduce the time for resolving the errors in ML model.
● The professional accountant should take effective measures such as reliable source of
information for cost saving and time saving.

Potential future uses

Developing innovation in accounting will be prompted by machine learning:

Over the years accounting firms are being adopted with the latest computing and
information systems (Marr, 2017). As a result, accounting firms and the occupation of
accountancy is more error free and transparent as well as fast. For instance, initially accounting
software companies used to provide desktop support to accounting firms, with time and the
developments of cloud-based services, now they offer cloud-based services to accounting firms.
Now, they are focusing on machine learning so that, accounting firms serve customers more
efficiently and provide more real-time insights with less spending. Hence, doing that machine
learning will play a vital role to achieve that promptness and efficiency.
Emerging Accounting Technologies - Machine Learning 15

Tasks that can be done by the machines with the help of Machine Learning:

There is a doubt that, with the developments of Machine Learning, the human workforce
will be impacted badly. This is because, there is a rumor that, machines will be replacing the
human workforce of accounting firms. That is totally not true. Machine Learning will be helping
the human working force in pair to avoid time consuming and repetitive less productive tasks.
This is how human workforce will be more concentrated on more critical accounting tasks and
innovation. Machine Learning can help accounting firms in many ways, followings are few to
mention:

Accounting firms expense claiming and submissions scrutinizing and audit:

With the help of machine learning, machines can help accounting firms with their
expenses claiming and scrutinizing. Machines can go through the company’s expense policies,
expense claims, fees structures matching, billing. Thus, when expenses/bills are claimed, then
machines can audit those documents and verify whether they are aligned with the company’s
expense policies. If otherwise, they can separate questionable claims/bills for further human
approval. Once, machines are set up for these tasks, they can handle large amount of these tasks,
saving human working hours.

Proper payments systems implementations and clearing bills against exact invoice:

Recently, billing systems are more complex than ever. Customers pay bills with multiples
invoices. As billing systems are more complex these days, thus many discrepancies and
mismatch occur (Marr, 2017). As a result, customers don not get update on their billing systems
and at the same time staffs at the accounting firms or accounting departments face immense
pressure to match the exact invoice which is very complex and time consuming. These issues can
be avoided with help of machine learning. Machines will be able to analyze the bills and invoices
and match based on the track and billing history of a customer. On the other hand, customers will
get the updated billing information. This way billing system can be improvised with the help of
machine learning and avoid unnecessary time consuming and increase efficiency.
Emerging Accounting Technologies - Machine Learning 16

Machine learning can help assess and avoid risks:

Machine learning can help accounting professionals in a great way to assess risks and
avoiding potential risks based on data. Based on existing data from previous projects completed
and data from on-going projects, machines can compare and assess potential risks and outline
tentative measures to avoid those risks. On the other hand, this will be way more time efficient
from human workforce within the same time frame. With these results human work force can
implement the solutions more promptly in the on-going projects and get the proper outcomes
from the projects.

Data analysis and calculation, performance reporting:

Within the context of data analysis, accounting departments and firms can be leveraged
by machine learning (Marr, 2017). Machines can learn and get updated with data from over the
years of activities. As a result, when accounting departments or firms are asked to give
performance report against any particular product or department or overall revenue, machines
can help answering these queries not letting human workforce to spend time doing all the
calculations and comparisons. On the other hand, machines can update with performance reports
so that, human workforce can keep themselves updated and optimize work progress.

Reconciliation tasks in the financial institutions can be a real-time task:

Reconciliation is a big task these days in the financial institutions. This is because it helps
financial institutions balancing their payables and receivables. It also key factor to customer
satisfactions. Especially, due to sharp rise of online banking recently, lots of disputes are
occurring thus, reconciliation is mandatory task to deliver excellent customer support and
eliminating disputes. Machines can help doing this task more efficiently and real-time basis and
can send the report to human workforce if any irregularities are caught. With the help of machine
learning machines can help balance financial institutions payables and receivables. This will not
only help keep the balances updated but also keeping the high customer service standard.
Emerging Accounting Technologies - Machine Learning 17

Case study
Tesla, Inc. is an American electric vehicle and clean energy organization which includes
the products like electric cars, solar products and many more related products. Tesla is using
machine learning for building one of the most innovative network across the world. The
company is analyzed as leading company across the world in providing the innovative and
creative products (Kim, et al., 2013). Their aim is to create the line of unique products and
making it affordable.

One of the unique features of Tesla's machine learning program is that the entire Tesla
fleet is a wireless network. When a car learns something new from other data inputs, every Tesla
car can immediately share the improvements. Tesla Motors is known for leading the self-driving
car revolution across the world. They includes to achieving high reliability in autonomous
vehicles without the use of LIDAR or HD maps. The company depends on the computer vision
entirely (Teece, 2018).

Tesla is a vertically integrated company, and the same is true for the intelligence of
autopilot. Everything that makes the Tesla Autopilot the best system in the world is based on
machine learning and raw video streams from 8 cameras around the vehicle. The company uses
the cameras footage in convolutional neutral networks for detecting the objects and taking the
action immediately.
Emerging Accounting Technologies - Machine Learning 18

In case of Machine learning framework, TensorFlow and PyTorch are used by the
company. ML is adopted for enhancing the business scalability and operations. Tesla is also
learning to deal with the data sets which are not approachable with the functions and costs. The
Autopilot software is a latest update which supports complex tasks like confirming automatic
lane changes by the driver and ramp transitions on highways (Mahajan, et al., 2016).

Tesla also uses the big data for fixing the problems which helps in enhancing the
customer satisfaction. The data is gathered from the online forum of the customers and they uses
the data for improving the next production. This helps the company in analyzing the requirement
of customers. Tesla uses the machine learning for analyzing the customer requirement uses the
gathered data which is collected from over 100 million miles with the autopilot software. They
compile the data in the cloud for generating the road map for automatic and driverless cars which
Tesla ensures that it is 100 times safe. It ensures planning and action, vehicle machine learning
algorithms must be able for predicting the results based on large amounts of data (Ganesh, 2017).
Tesla's self-driving car technology has not yet been up and running, but the best-case scenario
will even include the Internet of Things, where all self-driving cars are communicating some
basic details with each other.

Machine learning of Tesla effectively crowd sources some of the important data from the
vehicles and their drivers through internal and external sensors which helps in picking up the
information regarding the driver which helps in keeping the things operated. Machine learning in
Tesla includes the responsibility of taking care to provide the training to entire set of fleet and at
the individual level, the various computing devices are involved which helps in taking the
decisions. The decision making with ML also exists with the cars for forming the network with
other Tesla cars which is nearby for sharing the local insights as well as information.

Tesla use the machine learning for providing the safety to people. Tesla applies imitation
learning to driving tasks, such as how to deal with three-dimensional steep curves on highways,
or how to turn left at intersections. It helps in providing the accuracy as well as reducing the
various costs of the organization. Tesla ensure to develop and deploy the autonomy for scale.
Tesla is one of the most active developers in the market (Reed, et al., 2016). They are no
exception when developing their own chips. Artificial intelligence and machine learning enable
the company to design self-driving cars, which may change the way to drive cars. It can be said
Emerging Accounting Technologies - Machine Learning 19

with certainty that in the next few years, these AI chips will be further upgraded to promote this
movement automation journey.

The Tesla's exclusive infotainment system is analyzed as the lies in its user experience
and autonomous driving functions. Over the years, the company has performed many over-the-
air updates, and most other automakers around the world are not yet familiar with these updates.
Company employees will regularly upgrade and enhance the operating system of Tesla vehicles
(Howar, et al., 2016).

Tesla is a company which is dedicated to data collection and analysis. They are no
exception when developing their own chips. Machine learning and data analysis enable the
company to design self-driving cars, which may change the way customer drive cars. It does not
only include design and manufacturing, but the company also uses AI to process customer data
and even parse the online forums to get textual insights on common issues (Teece, 2018).
Emerging Accounting Technologies - Machine Learning 20

Conclusion
At the end of the report, it is analyzed that the Machine Learning technology, therefore,
happens to be the most essential technology having a tremendous amount of success rate. Present
day modern world requires modern solutions, such as automatically detecting an email from
spam folders or driving a car automatically. Human beings, in their day-to-day life face many
problems, which remain unsolved. ML in that case, can be handy, solving problems in a similar
way like a human brain does. It is analyzed that despite a lot of AI and Machine Learning pluses,
even financially strong businesses usually go through a difficult time obtaining the real value of
the technology. Accounting Profession desires to take advantage of distinctive prospects of
Machine Learning, but they have an imprecise idea about the use. While Machine Learning is
currently more prevalent compared to the previous 10 years, it is anticipated that ML will have
an impact on day-to-day activities and that business operations will soon expand significantly.
Global public accounting firms such as Price Water House Coopers are committed to supplying
their clients with the skills, they need to deploy Machine Learning Algorithms in organizations
to make better decisions while minimizing costs. The report has also included the case study of
Tesla which uses the machine learning in developing the products and providing the services to
customers.
Emerging Accounting Technologies - Machine Learning 21

Reference
Cawi, E., La Rosa, P.,S., & Nehorai, A. (2019). Designing machine learning workflows with an
application to topological data analysis. PLoS One, 14(12)
doi:http://dx.doi.org/10.1371/journal.pone.0225577
Chen, G., Li, S., Knibbs, L. D., Hamm, N. A., Cao, W., Li, T., ... & Guo, Y. (2018). A machine
learning method to estimate PM2. 5 concentrations across China with remote sensing,
meteorological and land use information. Science of the Total Environment, 636, 52-60.

Chicco, D. (2017). Ten quick tips for machine learning in computational biology. BioData
mining, 10(1), 35.

Fiorentini, N., & Losa, M. (2020). Handling imbalanced data in road crash severity prediction by
machine learning algorithms. Infrastructures, 5(7), 61.

Ganesh, M. I. (2017). Entanglement: Machine learning and human ethics in driver-less car
crashes. MACHINE RESEARCH, 76.

Haider, N., Baig, M. Z., & Imran, M. (2020). Artificial Intelligence and Machine Learning in 5G
Network Security: Opportunities, advantages, and future research trends. arXiv preprint
arXiv:2007.04490.

Howar, F., Meinke, K., & Rausch, A. (2016, October). Learning systems: machine-learning in
software products and learning-based analysis of software systems. In International
Symposium on Leveraging Applications of Formal Methods (pp. 651-654). Springer,
Cham.

Jiang, H., & Nachum, O. (2020, June). Identifying and correcting label bias in machine learning.
In International Conference on Artificial Intelligence and Statistics (pp. 702-712).

Kim, M., Wu, G., & Shen, D. (2013, September). Unsupervised deep learning for hippocampus
segmentation in 7.0 Tesla MR images. In International Workshop on Machine Learning
in Medical Imaging (pp. 1-8). Springer, Cham.
Emerging Accounting Technologies - Machine Learning 22

Kodelja, Z. (2019). Is machine learning real learning? CEPS Journal : Center for Educational
Policy Studies Journal, 9(3), 11-23. doi:http://dx.doi.org/10.26529/cepsj.709
Leavy, S. (2018, May). Gender bias in artificial intelligence: The need for diversity and gender
theory in machine learning. In Proceedings of the 1st international workshop on gender
equality in software engineering (pp. 14-16).

Lee, H., Huang, C., Yune, S., Tajmir, S. H., Kim, M., & Do, S. (2019). Machine friendly
machine learning: Interpretation of computed tomography without image
reconstruction. Scientific Reports (Nature Publisher Group), 9, 1-9.
doi:http://dx.doi.org/10.1038/s41598-019-51779-5
Mahajan, D., Park, J., Amaro, E., Sharma, H., Yazdanbakhsh, A., Kim, J. K., & Esmaeilzadeh,
H. (2016, March). Tabla: A unified template-based framework for accelerating statistical
machine learning. In 2016 IEEE International Symposium on High Performance
Computer Architecture (HPCA) (pp. 14-26). IEEE.

Marr, B. (2017, July 15). Machine Learning, Artificial Intelligence - And The Future Of
Accounting. Retrieved November 15, 2020, from
https://www.forbes.com/sites/bernardmarr/2017/07/07/machine-learning-artificial-
intelligence-and-the-future-of-accounting/?sh=500122492dd1

Ngiam, K. Y., & Khor, W. (2019). Big data and machine learning algorithms for health-care
delivery. The Lancet Oncology, 20(5), e262-e273.

Özer Çelik. (2018). A research on machine learning methods and its applications. Journal of
Educational Technology and Online Learning, 1(3), 25-40.
doi:http://dx.doi.org/10.31681/jetol.457046
Poolwan, J., & Smanchat, S. (2018). An architecture for simplified and automated machine
learning. International Journal of Electrical and Computer Engineering, 8(5), 2994-
3002. Retrieved from https://search.proquest.com/docview/2124487786?
accountid=188056
Reed, C., Kennedy, E., & Silva, S. (2016). Responsibility, Autonomy and Accountability: legal
liability for machine learning. Queen Mary School of Law Legal Studies Research Paper,
(243).
Emerging Accounting Technologies - Machine Learning 23

Stilgoe, J. (2018). Machine learning, social learning and the governance of self-driving
cars. Social studies of science, 48(1), 25-56.

Surender, R. S. (2020). A survey of big data and machine learning.  International Journal of
Electrical and Computer Engineering,  10(1), 575-580.
doi:http://dx.doi.org/10.11591/ijece.v10i1.pp575-580
Teece, D. J. (2018). Tesla and the reshaping of the auto industry. Management and Organization
Review, 14(3), 501-512.

Tiffin, P. A., & Paton, L. W. (2018). Rise of the machines? machine learning approaches and
mental health: Opportunities and challenges. The British Journal of Psychiatry, 213(3),
509-510. doi:http://dx.doi.org/10.1192/bjp.2018.105
Veale, M., & Binns, R. (2017). Fairer machine learning in the real world: Mitigating
discrimination without collecting sensitive data. Big Data & Society, 4(2),
2053951717743530.

Yeom, S., Giacomelli, I., Fredrikson, M., & Jha, S. (2018, July). Privacy risk in machine
learning: Analyzing the connection to overfitting. In 2018 IEEE 31st Computer Security
Foundations Symposium (CSF) (pp. 268-282). IEEE.

Richins, G., Stapleton, A., Stratopoulos, T. C., & Wong, C. (2017). Big data analytics:
Opportunity or threat for the accounting profession?. Journal of Information
Systems, 31(3), 63-79.

Stancheva-Todorova, E. P. (2018). How artificial intelligence is challenging accounting


profession. Journal of International Scientific Publications" Economy & Business, 12,
126-141.

Bakarich, K. M., & O'Brien, P. (2020). The Robots are Coming... But Aren't Here Yet: The Use
of Artificial Intelligence Technologies in the Public Accounting Profession. Journal of
Emerging Technologies in Accounting.

Cho, S., Vasarhelyi, M. A., Sun, T., & Zhang, C. (2020). Learning from Machine Learning in
Accounting and Assurance.
Emerging Accounting Technologies - Machine Learning 24

Haq, I., Abatemarco, M., & Hoops, J. (2020). The Development of Machine Learning and its
Implications for Public Accounting. The CPA Journal, 90(6), 6-9.

Bräuning, M., Hüllermeier, E., Keller, T., & Glaum, M. (2017). Lexicographic preferences for
predictive modeling of human decision making: A new machine learning method with an
application in accounting. European Journal of Operational Research, 258(1), 295-306.

You might also like