MACHINE

LEARNING

Abhayakrishna V V 19001
Ashwin Shetty 19013
Kattamuri Jayarajyalakshmi 19025
Prabjot Kaur 19037
Shraddha Tiwari 19049
Contents
Introduction to machine learning ............................................................................................... 2
Definition ................................................................................................................................... 2
Classification of machine learning............................................................................................. 2
Supervised machine learning: ............................................................................................. 2
Unsupervised machine learning: ........................................................................................ 2
Reinforcement machine learning: ...................................................................................... 3
Terminologies of Machine Learning ......................................................................................... 3
Applications of Machine Learning ............................................................................................ 3
Machine Learning Application in Automobile Industry ............................................................ 5
Tesla Autopilot........................................................................................................................... 6
Driving Features: ................................................................................................................. 6
Challenges faced by driverless cars .................................................................................... 6
Opportunities for machine learning: .......................................................................................... 8
Concerns about machine learning and its applications: ............................................................. 8
Other challenges in machine learning: ....................................................................................... 8
Bibliography .............................................................................................................................. 9

1
MACHINE LEARNING
Introduction to machine learning
Recent years have seen exciting advances in machine learning, which have raised its
capabilities across a suite of applications. Increasing data availability has allowed machine
learning systems to be trained on a large pool of examples, while increasing computer
processing power has supported the analytical capabilities of these systems. Within the field
itself there have also been algorithmic advances, which have given machine learning greater
power. As a result of these advances, systems which only a few years ago performed at
noticeably below-human levels can now outperform humans at some specific tasks.
Many people now interact with systems based on machine learning every day, for example in
image recognition systems, such as those used on social media; voice recognition systems, used
by virtual personal assistants; and recommender systems, such as those used by online retailers.
As the field develops further, machine learning shows promise of supporting potentially
transformative advances in a range of areas, and the social and economic opportunities which
follow are significant.
In healthcare, machine learning is creating systems that can help doctors give more accurate or
effective diagnoses for certain conditions. In transport, it is supporting the development of
autonomous vehicles, and helping to make existing transport networks more efficient. For
public services it has the potential to target support more effectively to those in need, or to
tailor services to users. And in science, machine learning is helping to make sense of the vast
amount of data available to researchers today, offering new insights into biology, physics,
medicine, the social sciences, and more.

Definition
Machine learning is a branch of artificial intelligence that allows computer systems to learn
directly from examples, data, and experience. Through enabling computers to perform specific
tasks intelligently, machine learning systems can carry out complex processes by learning from
data, rather than following pre-programmed rules.

Classification of machine learning

Supervised machine learning: It is by using labelled examples to predict future events.
Starting from the analysis of a known training dataset, the learning algorithm produces an
inferred function to make predictions about the output values. The system is able to provide
targets for any new input after sufficient training. The learning algorithm can also compare its
output with the correct, intended output and find errors in order to modify the model
accordingly.
The computer is presented with example inputs and their desired outputs, given by a “teacher”,
and the goal is to learn a general rule that maps inputs to outputs. The training process continues
until the model achieves the desired level of accuracy on the training data.
Unsupervised machine learning: It is used when the information used to train is neither
classified nor labelled. Unsupervised learning studies how systems can infer a function to
describe a hidden structure from unlabelled data. The system doesn’t figure out the right output,

2
but it explores the data and can draw inferences from datasets to describe hidden structures
from unlabelled data.
No labels are given to the learning algorithm, leaving it on its own to find structure in its input.
It is used for clustering population in different groups. Unsupervised learning can be a goal in
itself.
Reinforcement machine learning: It is a learning method that interacts with its environment
by producing actions and realizes errors or rewards. Trial and error search and delayed reward
are the most relevant characteristics of reinforcement learning. This method allows machines
and software agents to automatically determine the ideal behaviour within a specific context in
order to maximize its performance. Simple reward feedback is required for the agent to learn
which action is best; this is known as the reinforcement signal.
A computer program interacts with a dynamic environment in which it must perform a certain
goal (such as driving a vehicle or playing a game against an opponent). The program is
provided feedback in terms of rewards and punishments as it navigates its problem space.

Terminologies of Machine Learning

Model: A model is a specific representation learned from data by applying some machine
learning algorithm. A model is also called hypothesis.
Feature: A feature is an individual measurable property of our data. A set of numeric features
can be conveniently described by a feature vector. Feature vectors are fed as input to the model.
For example, in order to predict a fruit, there may be features like colour, smell, taste, etc.
Choosing informative, discriminating and independent features is a crucial step for effective
algorithms. We generally employ a feature extractor to extract the relevant features from the
raw data.
Target: A target variable or label is the value to be predicted by our model. For the fruit
example discussed in the features section, the label with each set of input would be the name
of the fruit like apple, orange, banana, etc.
Training: The idea is to give a set of inputs (features) and it’s expected outputs (labels), so
after training, we will have a model (hypothesis) that will then map new data to one of the
categories trained on.
Prediction: Once our model is ready, it can be fed a set of inputs to which it will provide a
predicted output(label).

Applications of Machine Learning

The value of machine learning technology has been recognized by companies across several
industries that deal with huge volumes of data. By leveraging insights obtained from this data,
companies are able work in an efficient manner to control costs as well as get an edge over
their competitors.
Financial Services: Companies in the financial sector are able to identify key insights in
financial data as well as prevent any occurrences of financial fraud, with the help of machine
learning technology. The technology is also used to identify opportunities for investments and

3
trade. Usage of cyber surveillance helps in identifying those individuals or institutions which
are prone to financial risk and take necessary actions in time to prevent fraud.
Marketing and Sales: Companies are using machine learning technology to analyse the
purchase history of their customers and make personalized product recommendations for their
next purchase. This ability to capture, analyse, and use customer data to provide a personalized
shopping experience is the future of sales and marketing.
Government: For government, machine learning offers the promise of more efficient and
effective services, through targeted interventions and tailoring of services. Government
agencies like utilities and public safety have a specific need for machine learning, as they have
multiple data sources, which can be mined for identifying useful patterns and insights. For
example: Sensor data can be analysed to identify ways to minimize costs and increase
efficiency. Furthermore, machine learning can also be used to minimize identity thefts and
detect fraud.
Healthcare: With the advent of wearable sensors and devices that use data to access health of
a patient in real time, machine learning is becoming a fast-growing trend in healthcare. Sensors
in wearable provide real-time patient information, such as overall health condition, heartbeat,
blood pressure and other vital parameters. Doctors and medical experts can use this information
to analyse the health condition of an individual, draw a pattern from the patient history, and
predict the occurrence of any complaints in the future. The technology also empowers medical
experts to analyse data to identify trends that facilitate better diagnoses and treatment.
Transportation: Based on the travel history and pattern of traveling across various routes,
machine learning can help transportation companies predict potential problems that could arise
on certain routes, and accordingly advise their customers to opt for a different route.
Transportation firms and delivery organizations are increasingly using machine learning
technology to carry out data analysis and data modelling to make informed decisions and help
their customers make smart decisions when they travel.
Energy: Machine learning can be used to optimise energy infrastructure. It can analyse patterns
of energy use and use these to design systems that can respond more effectively to peak
demands. For example, Google DeepMind has used machine learning to optimise the heating
and cooling requirements of its data centres, by predicting temperatures and pressures in the
data centre. DeepMind’s algorithm was able to reduce the amount of energy needed for cooling
by 40%. Having tested the system in a live data centre, DeepMind plans to roll it out more
broadly. There may be future applications for a similar system in improving the efficiency of
power plants.

Oil and Gas: This is perhaps the industry that needs the application of machine learning the
most. Right from analysing underground minerals and finding new energy sources to streaming
oil distribution, machine learning applications for this industry are vast and are still expanding.
Pharmaceuticals:
The pharmaceuticals sector both relies on and creates large amounts of data, from clinical trials,
from drug efficacy studies, or from genetic studies. These large-scale datasets require methods
to aid their analysis, in order to extract valuable insights that can improve research and
development processes, and to create diagnostic tools to target medicines at patients who will
most benefit.

4
Social Media Services: From personalizing your news feed to better ads targeting, social media
platforms are utilizing machine learning for their own and user benefits. Here are a few
examples that you must be noticing, using, and loving in your social media accounts, without
realizing that these wonderful features are nothing but the applications of machine learning.
People you may know: Machine learning works on a simple concept: understanding with
experiences. Facebook continuously notices the friends that you connect with, the profiles that
you visit very often, your interests, workplace, or a group that you share with someone etc. On
the basis of continuous learning, a list of Facebook users are suggested that you can become
friends with.
Facial Recognition: You upload a picture of you with a friend and Facebook instantly
recognizes that friend. Facebook checks the poses and projections in the picture, notice the
unique features, and then match them with the people in your friend list. The entire process at
the backend is complicated and takes care of the precision factor but seems to be a simple
application of ML at the front end.
Similar Pins: Machine learning is the core element of Computer Vision, which is a technique
to extract useful information from images and videos. Pinterest uses computer vision to identify
the objects (or pins) in the images and recommend similar pins accordingly.

Machine Learning Application in Automobile Industry

Machine Learning applications include evaluation of driver condition or driving scenario
classification through data fusion from different external and internal sensors. We examine
different algorithms used for self-driving cars.
Today, the machine learning algorithms are extensively used to find the solutions to various
challenges arising in manufacturing self-driving cars. With the incorporation of sensor data
processing in an ECU (Electronic Control Unit) in a car, it is essential to enhance the utilization
of machine learning to accomplish new tasks. The potential applications include evaluation of
driver condition or driving scenario classification through data fusion from different external
and internal sensors – like lidar, radars, cameras or the IoT (Internet of Things).
The applications that run the information system of a car can receive the information from
sensor data fusion systems and for example, have the capability to direct the car to a hospital
if it notices that something is not right with the driver. This application based on machine
learning also includes the driver’s speech and gesture recognition and language translation. The
algorithms are classified as unsupervised and supervised algorithms. The difference between
both of them is how they learn.
In the autonomous car, one of the major tasks of a machine learning algorithm is continuous
rendering of surrounding environment and forecasting the changes that are possible to these
surroundings. These tasks are classified into 3 sub-tasks:
 The detection of an Object: Proximity sensors are used for detection or surrounding
objects. Proximity sensors can have a high reliability and long functional life because
of the absence of mechanical parts and lack of physical contact between the sensor and
the sensed object.
 The Identification of an Object or recognition object classification: Identification of
objects is done by inbuild programs.
5
  The Object Localization and Prediction of Movement: A computer program interacts
with processed input given by proximity sensors. Based on the information received,
the braking system and steering systems of the vehicle respond.

Tesla Autopilot
Tesla Autopilot is an advanced driver-assistance system feature offered by Tesla that has lane
centring, adaptive cruise control, self-parking, the ability to automatically change lanes, and
the ability to summon the car to and from a garage or parking spot.
Driving Features:
Adaptive cruise control: Autopilot has the ability to follow another car, maintaining a safe
distance from it as it speeds up and slows down. It can observe a second vehicle in front of the
vehicle that it is following as well as differentiate between pedestrians, bicyclists/motorcyclists,
small cars, and large SUVs/trucks. It also slows on tight curves, on interstate ramps, and when
a car crosses the road in front of it. It can be enabled at any speed between 18 mph and 90 mph.
By default, it sets the limit at the current speed limit plus/minus any driver-specified offset,
then adjusting speed according to changes in speed limits.
Alerts: Autopilot alerts the driver under various circumstances, such as a surprising situation
on the road or excessive inattention by the driver. If the driver dismisses three audio warnings
within an hour, Autopilot is disabled until the car is parked. This is to prevent experienced
drivers from excessive reliance on built-in safety features. At speeds under 8 mph on divided
highways, Autopilot functions indefinitely without the driver's hands on the wheel. Under
45 mph free hands are allowed for five minutes, unless the car detects lateral acceleration.
Above 45 mph free hands are allowed for three minutes if following another vehicle or one
minute without following a car.
Auto park: Auto park drives the car into a parking spot, while Summon drives it out.
Configuration settings control maximum distance, side clearance and bumper clearance. This
feature activates Homelink to open and close garage doors and it is available using the fob or
the Tesla mobile app.
Auto Steering: Autosteer controls the position of the car to remain in whatever lane it is in (known as
lane-keeping).
Safety features: The Autopilot can detect a potential front or side collision with another
vehicle, bicycle or pedestrian within a distance of 525 feet (160 m), if one is found it sounds a
warning. Autopilot has automatic emergency braking that detects objects that may hit the car
and applies the brakes. The car may also automatically swerve out of the way to prevent fast
moving collisions. Autopilot also can automatically adjust the high/low beam headlights.
Challenges faced by driverless cars
Road conditions: Road conditions could be highly unpredictable and vary from places to
places. In some cases, there are smooth and marked broad highways. In other cases, road
conditions are highly deteriorated - no lane marking. Lanes are not defined, there are potholes,
mountainous and tunnel roads where external signals for direction are not very clear and
likewise.

6
Weather conditions: Weather conditions play another spoilsport. There could be a sunny and
clear weather or rainy and stormy weather. Autonomous cars should work in all sorts of weather
conditions. There is absolutely no scope for failure or downtime.
Traffic Conditions: Autonomous cars would have to get onto the road where they would have
to drive in all sorts of traffic conditions. They would have to drive with other autonomous cars
on the road, and at the same time, there would also be a lot of humans. Wherever humans are
involved, there are involved a lot of emotions. Traffic could be highly moderated and self-
regulated. But often there are cases where people may be breaking traffic rules. An object may
turn up in unexpected conditions. In the case of dense traffic, even the movement of few cms
per minute does matter. One can’t wait endlessly for traffic to automatically clear and have
some precondition to start moving. If more of such cars on the road are waiting for traffic to
get cleared, ultimately that may result in a traffic deadlock.
Accident liability: The most important aspect of autonomous cars is accidents liability. Who
is liable for accidents caused by a self-driving car? In the case of autonomous cars, the software
will be the main component that will drive the car and will make all the important decisions.
While the initial designs have a person physically placed behind the steering wheel, newer
designs showcased by Google, do not have a dashboard and a steering wheel! In such designs,
where the car does not have any controls like a steering wheel, a brake pedal, an accelerator
pedal, how is the person in the car supposed to control the car in case of an untoward incident?
Additionally, due to the nature of autonomous cars, the occupants will mostly be in a relaxed
state and may not be paying close attention to the traffic conditions. In situations where their
attention is needed, by the time they need to act, it may be too late to avert the situation.
Radar interference: Autonomous cars use lasers and radar for navigation. The lasers are
mounted on roof top while the sensors are mounted on the body of the vehicle. The principle
of radar works by detecting reflections of radio waves from surrounding objects. When on the
road, a car will continuously emit radio frequency waves, which get reflected from the
surrounding cars and other objects near the road. The time taken for the reflection is measured
to calculate the distance between the car and the object. Appropriate action is then taken based
on the radar readings. The principle of radar works by detecting reflections of radio waves from
surrounding objects. When on the road, a car will continuously emit radio frequency waves,
which get reflected from the surrounding cars and other objects near the road. The time taken
for the reflection is measured to calculate the distance between the car and the object.
Appropriate action is then taken based on the radar readings. When this technology is used for
hundreds of vehicles on the road, will a car be able to distinguish between its own (reflected)
signal and the signal (reflected or transmitted) from another vehicle? Even if multiple radio
frequencies are available for radar, this frequency range is unlikely to be insufficient for all the
vehicles manufactured.

7
Opportunities for machine learning:
The significant potential of machine learning was clear to many, not least because of its
connection to the world of ‘big data’ and its ability to analyse data. By analysing this data,
participants thought that machine
learning could:
 Be more objective than human users, or help avoid cases of human error, for example
avoiding issues that may arise where decision-makers are tired or emotional.
 Be more accurate, for example in detecting features in medical images and making
accurate diagnoses.
 Be more efficient, particularly in terms of public sector resources and shaping how
services were delivered.
 Offer opportunities for new businesses, and economic growth across a range of sectors.
 Play a role in addressing large-scale societal challenges, such as climate change or the
pressures of an aging population.

Concerns about machine learning and its applications:

 The potential for machine learning systems to cause harm, for example as a result of
accidents in autonomous vehicles.
 The possibility that people could be replaced by machines in the workplace, or could
become over-reliant on machines, for example in making diagnoses.
 The extent to which systems using machine learning might make experiences less
personal, or human, either by changing the nature of valued activities, or by making
generalised predictions about an individual.
 The idea that machine learning systems could restrict the choices open to an individual,
for example directing consumers to one type of product or service.

Other challenges in machine learning:

 Machine learning could generate an algorithmic bubble, in which unusual or
challenging opinions, experiences or interactions were filtered-out, ultimately
narrowing the horizons of its users.
 People could be mis-labelled, or inadvertently stereotyped, and have their activities
mistakenly restricted as a result.
 Machine learning further destabilises the current distinction between ‘sensitive’ or
‘personal’ and ‘non-sensitive’ data, it allows datasets which at first seem innocuous to
be employed in ways that allow the sensitive to be inferred from the mundane.
 Main concern is about privacy and data use is variable both between individuals and
for each individual between applications.

8
Bibliography
 Retrieved from https://www.kdnuggets.com/2017/06/machine-learning-algorithms-
used-self-driving-cars.html
 Retrieved from https://en.wikipedia.org/wiki/Tesla_Autopilot
 Retrieved from https://iiot-world.com/artificial-intelligence/five-challenges-in-
designing-a-fully-autonomous-system-for-driverless-cars/
 Retrieved from https://www.expertsystem.com/machine-learning-definition/
 Retrieved from https://www.geeksforgeeks.org/getting-started-machine-learning/
 Retrieved from https://royalsociety.org/topics-policy/projects/machine-learning/
 Reference https://www.tesla.com/autopilot
 Reference source https://www.mckinsey.com/industries/high-tech/our-insights/an-
executives-guide-to-machine-learning

9
10