Applications of AI, ML, DL

Dr. Kavita Mohite (Bhosle)

 Human Resource
Advanced Healthcare Lifestyle
Analysis & Visualization Social media
(Computer Vision) Gaming
Predictions Astronomy
NLP Chatbots
Artificial Intelligence in Surveillance
E-Commerce Finance
AI in Education Data Security
Purpose Travel and
Artificial Intelligence in Transport
Robotics Marketing
GPS and Navigations Entertainment
Healthcare Military
Automobiles Speech
Agriculture Recognition
Expert system
1. AI Application in E-Commerce
Personalized Shopping, AI-Powered Assistants, Fraud Prevention
2. Applications Of Artificial Intelligence in Education
Administrative Tasks Automated to Aid Educators, Creating Smart
Content
Voice Assistants, Personalized Learning
3. Applications of Artificial Intelligence in Lifestyle
Autonomous Vehicles, Spam Filters, Facial Recognition,
Recommendation System, Shopping Experience, Language
Translation
4. Applications of Artificial Intelligence in Navigation
GPS technology can provide users with accurate, timely, and
detailed information to improve safety, Traffic Prediction,
Positioning & Planning, Personalization (Intelligent Routing)
5. Artificial Intelligence in Robotics
NLP, Object Recognition & Manipulation, Human-Robotics
Interaction
6. Healthcare
Insights & Analysis, Telehealth, Patient Monitoring, Surgical
Assistance
7. Agriculture
Stock Monitoring, Supply Chain, Pest Management, Forecasting
8. Human Resource
Screening, Onboarding, Performance, Workforce Planning
9. Social media
Fraud Detection, Sentiment Analysis,
10. Gaming
Game Assistance, Animation
11. Astronomy
Detection & Classification, Analysis
Machine learning (ML)

ML is a subdomain of artificial intelligence (AI) that focuses on

developing systems that learn or improve performance based
on the data they ingest.

Machine learning is focuses on the use of data and algorithms to

copy the way that humans learn, gradually improving its
accuracy.

As it is evident from the name, it gives the computer that makes

it more similar to humans: The ability to learn.
Inductive Learning
An technique of machine learning called inductive learning
trains a model to generate predictions based on examples or
observations.

During inductive learning, the model picks up knowledge

from particular examples or instances and generalizes it
such that it can predict outcomes for new data.

In supervised learning situations, where the model is trained

using labeled data, inductive learning is frequently utilized.

Inductive learning is used by a number of well-known

machine learning algorithms, such as decision trees, k-
nearest neighbors, and neural networks.
Advantages
1. Inductive learning models are flexible and adaptive, they are well
suited for handling difficult, complex, and dynamic information.

2. Finding hidden patterns and relationships in data: Inductive

learning models are ideally suited for tasks like pattern recognition
and classification because they can identify links and patterns in data
that may not be immediately apparent to humans.

3. Huge datasets − Inductive learning models are suitable for

applications requiring the processing of massive quantities of data
because they can efficiently handle enormous volumes of data.

4. Appropriate for situations where the rules are ambiguous − Since

inductive learning models may learn from examples without explicit
programming, they are suitable for situations when the rules are not
precisely described or understood beforehand.
Disadvantages
1. May overfit to particular data − Inductive learning models that have
overfit to specific training data, or that have learned the noise in the
data rather than the underlying patterns, may perform badly on fresh
data.

2. Computationally costly possible − The employment of inductive

learning models in real-time applications may be constrained by their
computationally costly nature, especially for complex datasets.

3. Limited interpretability − Inductive learning models may be difficult to

understand, making it difficult to understand how they arrive at their
predictions, in applications where the decision-making process must be
transparent and explicable.

4. Inductive learning models are only as good as the data they are
trained on, therefore if the data is inaccurate or inadequate, the model
may not perform effectively.
Deductive Learning
Deductive learning is a method of machine learning in which a
model is built using a series of logical principles and steps.

In deductive learning, the model is specifically designed to

adhere to a set of guidelines and processes in order to
produce predictions based on new, unexplored data.

Deductive learning is used by a number of well-known

machine learning algorithms, such as decision trees, rule-
based systems, and expert systems.
Advantages
More effective − Since deductive learning begins with broad concepts
and applies them to particular cases, it is frequently quicker than
inductive learning.
Deductive learning can sometimes yield more accurate findings than
inductive learning since it starts with certain principles and applies them
to the data.
Deductive learning is more practical when data are sparse or challenging
to collect since it requires fewer data than inductive learning.
Disadvantages
Deductive learning is constrained by the rules that are currently in place,
which may be insufficient or obsolete.
Deductive learning is not appropriate for complicated issues that lack
precise rules or correlations between variables, nor is it appropriate for
ambiguous problems.
Results that are biased − The quality of the rules and knowledge base,
which might add biases and mistakes to the results, determines how
accurate deductive learning is.
 Inductive Learning Deductive Learning
Approach Bottom-up Top-down
Data Specific examples Logical rules and procedures
Model Creation Find correlations and patterns in obey clearly stated guidelines and
data. instructions
Training Adapting model parameters and Programming explicitly and
learning from instances establishing rules
Goal Using fresh data, generalizing, Make a model that precisely
and making predictions. complies with the given guidelines
and instructions.
Examples Decision trees, neural networks, Knowledge-based systems, expert
clustering algorithms systems, and rule-based systems
Strengths capable of learning from a variety accurately when according to
of complicated data, adaptable, established norms and processes,
and versatile and effective when doing specific
duties
Limitations It may be difficult to manage limited to well-defined duties and
complex and diverse data and norms, possibly incapable of
may overfit to specific facts. adjusting to novel circumstances
Types of Machine Learning Algorithm:-
Supervised learning
Unsupervised learning

Supervised Learning
The supervised learning approach of machine learning is the approach
with which the algorithms are trained by using labeled datasets.

A dataset is the collection of related yet discrete data, which can be

used or managed individually as well as a group.

The labelled datasets are the named pieces of data that are tagged
with one or multiple labels pertaining to certain properties or
characteristics.

The labeled datasets make the algorithms understand the relationship

among the datasets and carry out classification or prediction.
Different Approaches of Supervised Learning
Supervised learning is divided further into two approaches −
Classification − In this approach, algorithms are trained to
categorize the data into distinct units depending on their labels.
Examples of some classification algorithms are − Decision Tree,
Random Forest, Support Vector Machine, etc. Classification can be
of types Binary and Multi-class.
Regression − This approach makes a computer program understand
the relationship between dependent and independent data. As the
name suggests regression means "going back to", the algorithm is
exposed to the past data. Once training the algorithm is completed,
the algorithm can predict the future values easily. Some popular
regression algorithms are Linear, Logistic, and Polynomial
regression. Regression can be of types Linear and Non-linear.
The prominent difference between Classification and
Regression algorithms is that the Regression algorithms
are used to predict continuous values such as height,
weight, cost, salary, weather, etc. In contrast, the
Classification algorithms are used to classify or predict
discrete values such as True or False, Valid or Invalid, Yes
or No, Spam or Not Spam, Male or Female, etc.

Unsupervised Learning
The unsupervised learning approach of machine learning
does not use labelled datasets for training the algorithms.
Instead, the machines learn on their own by accessing
massive amount of unclassified data and finding its
implicit patterns. The algorithms analyze and cluster the
unlabelled datasets.
Different Approaches of Unsupervised Learning
The unsupervised learning approach is of the following three types

Association −This approach uses some rules to find relationships

between variables in a dataset. This approach is often used in
suggestions and recommendation. For example, suggesting an item
to a customer with: "The customers who bought this item also
bought", or "You may also like", or simply by showing allied product
images and recommending to buy related items. For example,
when the primary product being purchased is a computer, then
suggesting to buy a wireless mouse and a remote keyboard too.

Clustering − It is a learning technique in data mining where

unlabelled or unclassified data are grouped depending on either
similarities or differences among them. This technique is helpful for
the businesses to understand market segments depending on the
customers demographics.
Dimensionality Reduction − It is a learning technique used to
reduce the number of random variables or ‘dimensions’ to
obtain a set of principal variables, when the number of
variables is very high. This technique helps data compression
without compromising the usability of the data. This learning
is used for pre-processing of the audio/visual data to improve
the quality of the outcome or making the background of an
image transparent.

Unsupervised Learning Essential?

Unsupervised learning is essential because of the following
reasons −
Unlabelled, uncategorized data is available in abundance.
Unsupervised learning can explore unknown patterns of data.
Labelling the data is a tedious task, which also can allow
human errors in Supervised learning, which is not the case
with Unsupervised learning.
Factor Supervised Learning Unsupervised Learning
To train the algorithm for prediction. The To train the algorithm to find insights from the large
Objective outcome the algorithm predicts mostly volume of unclassified data.
occurs as per the human expectation.

The datasets used in Supervised learning The data used in Unsupervised learning are
Dataset Labelling
are labelled. unclassified.

Knowledge of Classes The classes of data are known. The number of classes is unknown as the model data
is uncategorized and unlabelled.

In supervised learning, human The unsupervised learning makes the algorithm to

intervention is required to label the data take care of both; the input and the output of the
Human Intervention
appropriately. data analizing but human intervention is only
required for data validation.

With remarkable amount of human With the less amount of human intervention,
Proximity with Artificial
intervention, Supervised learning seems Unsupervised learning is very close to Artificial
Intelligence
distant from the real Artificial Intelligence. Intelligence.

It is simple and inexpensive. It is complicated, timeconsuming, and requires more

Computational Complexity
resources.

In Supervised learning, the process of In case of unsupervised learning, the process of

Learning Process
training the algorithm takes place offline. training the algorithms takes place in real time.

It provides highly accurate outcome. The Unsupervised learning is less accurate.

accuracy can be hampered only if the
Accuracy of the Outcome
experts who are labelling the datasets
didn’t label them appropriately.
Reinforcement learning methods
In reinforcement learning methods, a trained agent interacts with a
specific environment and takes actions based upon the current
state of that environment.
The working of reinforcement learning is as follows −
First you need to prepare an agent with some specific set of
strategies.
Now leave the agent to observe the current state of the
environment.
Based on the agent's observation, select the optimal policy, and
perform suitable action.
Based on the action taken, the agent will get reward or penalty.
Update the set of strategies used in step 1, if needed. Repeat the
process from step1-4 until the agent learns and adopts the optimal
policy as well.
Iris Flowers Dataset
The Iris Flowers Dataset involves predicting the flower species given
measurements of iris flowers.
It is a multi-class classification problem. The number of observations for
each class is balanced. There are 150 observations with 4 input variables
and 1 output variable. The variable names are as follows:
1.Sepal length in cm.
2.Sepal width in cm.
3.Petal length in cm.
4.Petal width in cm.
5.Class (Iris Setosa, Iris Versicolour, Iris Virginica).
The baseline performance of predicting the most prevalent class is a
classification accuracy of approximately 26%.
A sample of the first 5 rows is listed below.

1 5.1,3.5,1.4,0.2,Iris-setosa
2 4.9,3.0,1.4,0.2,Iris-setosa
3 4.7,3.2,1.3,0.2,Iris-setosa
4 4.6,3.1,1.5,0.2,Iris-setosa
5 5.0,3.6,1.4,0.2,Iris-setosa
Swedish Auto Insurance Dataset
The Swedish Auto Insurance Dataset involves predicting the total
payment for all claims in thousands of Swedish Kronor, given the total
number of claims.
It is a regression problem. It is comprised of 63 observations with 1
input variable and one output variable. The variable names are as
follows:
1.Number of claims.
2.Total payment for all claims in thousands of Swedish Kronor.
The baseline performance of predicting the mean value is an RMSE of
approximately 81 thousand Kronor.
A sample of the first 5 rows is listed below.
1 108,392.5
2 19,46.2
3 13,15.7
4 124,422.2
5 40,119.4
https://archive.ics.uci.edu/
Applications of Deep Learning
Optical Character Recognition (OCR)
 Speech Recognition

Hinton et al. (2012), Graves et al. (2013), Chorowski et

al. (2015), Sak et al. (2015)
 Conversation Modeling

Shang et al. (2015), Vinyals et al. (2015), Weston et al.

(2016), Serban et al. (2017)
 Question Answering

Hermann et al. (2015), Chen et al. (2016), Wang et al.

(2017), Hu et al. (2017)
 Object Detection / Recognition

Long et al. (2015), Liang et al. (2015), He et al. (2017),

Caelles et al. (2017)
Vision as a source of semantic information

slide credit: Fei-Fei, Fergus & T

 Object categorization

sky
building

flag

face
banner
wall
street lamp
bus bus

cars slide credit: Fei-Fei, Fergus & T

Object Recognition in Mobile Phone
 Visual Tracking

Choi et al. (2017), Yun et al. (2017), Alahi et al. (2017)

Face Detection
Login without password
 Image Captioning

Mao et al. (2014), Kiros et al. (2015), Karpathy et al.

(2015), Chen et al. (2015)
 Video Captioning

Donahua et al. (2014), Venugopalan et al. (2014), Zhu

et al. (2015), Cho et al. (2015)
 Visual Question Answering

Santoro et al. (2017), Kazemi et al. (2016), Huet al.

(2017), Johnson et al. (2017)
 Video Summarization

Chheng et al. (2007), Ajmal et al. (2012), Zhang et al.

(2016), Zhong et al. (2017), Panda et al. (2017)
Google Car
Industrial Robot
Mobile Robot
Medical Imaging
 Generating Authentic Photos
Kingma et al. (2013), Goodfellow et al. (2014),Nguyen
et al. (2016), Karras et al. (2017)
 Generating Raw Audio

Oord et al. (2016)

 Errors in Machine Learning
•Reducible errors: These errors can be reduced to improve the model
accuracy. Such errors can further be classified into bias and Variance.

•Irreducible errors: These errors will always be present in the model

regardless of which algorithm has been used. The cause of these errors is
unknown variables whose value can't be reduced.
Bias
Bias refers to the difference between the expected value of the
model’s predictions and the true values of the target variable.
While making predictions, a difference occurs between prediction
values made by the model and actual values/expected values, and
this difference is known as bias errors or Errors due to bias.
Low Bias: A low bias model will make fewer assumptions about the
form of the target function.
High Bias: A model with a high bias makes more assumptions, and
the model becomes unable to capture the important features of
our dataset. A high bias model also cannot perform well on new
data.
How to reduce High Bias:
High bias mainly occurs due to a much simple model. Below are some ways to reduce
the high bias:
Increase the input features as the model is underfitted.
Decrease the regularization term.
Use more complex models, such as including some polynomial features.
The variance would specify the amount of variation in the
prediction if the different training data was used.

Low variance means there is a small variation in the

prediction of the target function with changes in the training
data set. At the same time, High variance shows a large
variation in the prediction of the target function with
changes in the training dataset.

A high variance model leads to overfitting.

Increase model complexities.
How to Reduce High Variance:

Reduce the input features or number of parameters as a

model is overfitted.

Do not use a much complex model.

Increase the training data.

Increase the Regularization term.

Low-Bias, Low-Variance:
The combination of low bias and low variance shows an ideal
machine learning model. However, it is not possible
practically.
Low-Bias, High-Variance: With low bias and high variance,
model predictions are inconsistent and accurate on average.
This case occurs when the model learns with a large number
of parameters and hence leads to an overfitting
High-Bias, Low-Variance: With High bias and low variance,
predictions are consistent but inaccurate on average. This case
occurs when a model does not learn well with the training
dataset or uses few numbers of the parameter. It leads
to underfitting problems in the model.
High-Bias, High-Variance:
With high bias and high variance, predictions are inconsistent
and also inaccurate on average.
Overfitting
Overfitting occurs when our machine learning model
tries to cover all the data points or more than the
required data points present in the given dataset.
Because of this, the model starts caching noise and
inaccurate values present in the dataset, and all these
factors reduce the efficiency and accuracy of the model.
The overfitted model has low bias and high variance.
To avoid the Overfitting in Model

Both overfitting and underfitting cause the degraded

performance of the machine learning model. But the
main cause is overfitting, so there are some ways by
which we can reduce the occurrence of overfitting in
our model.

1. Cross-Validation
2. Training with more data
3. Removing features
4. Early stopping the training
5. Regularization
6. Ensembling
Cross validation is a technique used in machine
learning to evaluate the performance of a model on
unseen data. It involves dividing the available data
into multiple folds or subsets, using one of these
folds as a validation set, and training the model on
the remaining folds.
Types of Cross-Validation
Leave-one-out cross validation
perform training on the whole dataset but leaves
only one data-point of the available dataset and
then iterates for each data-point. In LOOCV, the
model is trained on samples and tested on the one
omitted sample, repeating this process for each data
point in the dataset.
Holdout validation
Perform training on the 50% of the given dataset
and rest 50% is used for the testing purpose.

Stratified Cross-Validation
Ensure that each fold of the cross-validation
process maintains the same class distribution as the
entire dataset. This is particularly important when
dealing with imbalanced datasets, where certain
classes may be underrepresented
 k-fold cross validation
split the dataset into k number of subsets (known as folds)
then we perform training on the all the subsets but leave
one(k-1) subset for the evaluation of the trained model. In
this method, we iterate k times with a different subset
reserved for testing purpose each time.
Underfitting

Underfitting occurs when our machine learning model

is not able to capture the underlying trend of the
data.
An underfitted model has high bias and low variance.
To avoid underfitting:

1. By increasing the training time of the

model.
2. By increasing the number of features.
 Bias Variance tradeoff
The bias-variance tradeoff is a fundamental
concept in machine learning that describes
the relationship between a model’s
complexity, the accuracy of its predictions,
and how well it can make predictions on
previously unseen data that were not used to
train the model.

In essence, it is a tradeoff between the

model’s ability to minimize bias and variance
errors.
1. High bias models are typically too simple and fail to
capture the underlying patterns in the data
2. Variance, on the other hand, refers to the variability of
the model’s predictions for different training sets.
3. High variance models are typically too complex and
overfit the training data, resulting in poor generalization
performance
4. The goal is to find a model that strikes a balance between
bias and variance, which can be achieved by tuning the
model’s complexity or regularization parameters
5. In summary, the bias-variance tradeoff is a crucial
concept in machine learning that helps us understand the
relationship between a model’s complexity, its ability to
generalize to new data, and its prediction accuracy
Advantages of Deep Learning:
Deep learning has several advantages over traditional machine learning
methods, some of the main ones include:
Automatic feature learning: Deep learning algorithms can automatically learn
features from the data, which means that they don’t require the features to
be hand-engineered. This is particularly useful for tasks where the features are
difficult to define, such as image recognition.
Handling large and complex data: Deep learning algorithms can handle large
and complex datasets that would be difficult for traditional machine learning
algorithms to process. This makes it a useful tool for extracting insights from
big data.
Improved performance: Deep learning algorithms have been shown to
achieve state-of-the-art performance on a wide range of problems, including
image and speech recognition, natural language processing, and computer
vision.
Handling non-linear relationships: Deep learning can uncover non-linear
relationships in data that would be difficult to detect through traditional
methods.
Handling structured and unstructured data: Deep learning algorithms can
handle both structured and unstructured data such as images, text, and audio.
Predictive modeling: Deep learning can be used to make predictions
about future events or trends, which can help organizations plan for the
future and make strategic decisions.
Handling missing data: Deep learning algorithms can handle missing
data and still make predictions, which is useful in real-world
applications where data is often incomplete.
Handling sequential data: Deep learning algorithms such as Recurrent
Neural Networks (RNNs) and Long Short-term Memory (LSTM)
networks are particularly suited to handle sequential data such as time
series, speech, and text. These algorithms have the ability to maintain
context and memory over time, which allows them to make predictions
or decisions based on past inputs.
Scalability: Deep learning models can be easily scaled to handle an
increasing amount of data and can be deployed on cloud platforms and
edge devices.
Generalization: Deep learning models can generalize well to new
situations or contexts, as they are able to learn abstract and hierarchical
representations of the data.
Disadvantages of Deep Learning:
While deep learning has many advantages, there are also some
disadvantages to consider:
High computational cost: Training deep learning models requires
significant computational resources, including powerful GPUs and large
amounts of memory. This can be costly and time-consuming.
Overfitting: Overfitting occurs when a model is trained too well on the
training data and performs poorly on new, unseen data. This is a
common problem in deep learning, especially with large neural
networks, and can be caused by a lack of data, a complex model, or a
lack of regularization.
Lack of interpretability: Deep learning models, especially those with
many layers, can be complex and difficult to interpret. This can make it
difficult to understand how the model is making predictions and to
identify any errors or biases in the model.
Dependence on data quality: Deep learning algorithms rely on the
quality of the data they are trained on. If the data is noisy, incomplete,
or biased, the model’s performance will be negatively affected.
Data privacy and security concerns: As deep learning models often rely
on large amounts of data, there are concerns about data privacy and
security. Misuse of data by malicious actors can lead to serious
consequences like identity theft, financial loss and invasion of privacy.
Lack of domain expertise: Deep learning requires a good understanding
of the domain and the problem you are trying to solve. If the domain
expertise is lacking, it can be difficult to formulate the problem and
select the appropriate algorithm.
Unforeseen consequences: Deep learning models can lead to
unintended consequences, for example, a biased model can
discriminate against certain groups of people, leading to ethical
concerns.
Limited to the data its trained on: Deep learning models can only make
predictions based on the data it has been trained on. They may not be
able to generalize to new situations or contexts that were not
represented in the training data.
Black box models: some deep learning models are considered as “black-
box” models, as it is difficult to understand how the model is making
predictions and identifying the factors that influence the predictions.
Thank you