Machine Learning an Overview

Difference between AI, ML & DL

 AI simulates human intelligence to perform tasks and make
decisions.
 ML is a subset of AI that uses algorithms to learn patterns from
data.
 DL is a subset of ML that employs artificial neural networks for
complex tasks
Machine Learning
 Machine learning (ML) is defined as a discipline of artificial
intelligence (AI) that provides machines the ability to
automatically learn from data and past experiences to identify
patterns and make predictions with minimal human
intervention
 In general a training system to learn from past experiences and
improve performance over time.
 Deliver fast and accurate results to get profitable opportunities.
Real World Applications of Machine Learning
Real World Applications of Machine Learning
 Image Recognition
 Identify objects, persons, places, digital images, and so on
 Facebook provides us a feature of auto friend tagging suggestion - uses machine
learning's face detection and recognition algorithm.
 Speech recognition
 Speech to text
 E.g. Google assistant, Siri and Alexa
 Traffic Prediction
 Google Maps, shows us the correct path with the shortest route and predicts the traffic
conditions.
 Real Time location of the vehicle form Google Map app and sensors
 Average time has taken on past days at the same time.
 Product Recommendations
 e-commerce and entertainment companies such as Amazon, Netflix, etc
 Search for some product, an advertisement for the same product will appear while internet
surfing
 Self Driving Cars
 Tesla, car manufacturing company uses unsupervised learning method to train the car
Real World Applications of Machine Learning
 Email Spam and Malware filtering
 filtered automatically as important, normal, and spam
 Virtual Personal Assistant
 Through our voice instructions such as Play music, call someone, Open an email,
Scheduling an appointment, etc.
 Online Fraud Detection
 fake accounts, fake ids, and steal money in the middle of a transaction
 For each genuine transaction, there is a specific pattern which gets change for the fraud
transaction hence, it detects it and makes our online transactions more secure.
 Stock Market Trading
 long short term memory neural network- working with sequential data, which can vary in
length, and learning long-term dependencies between time steps of that data
 Medical Diagonosis
 Able to build 3D models and finds brain tumors and other brain-related diseases easily
 Automatic Language Translation
 Google's GNMT (Google Neural Machine Translation) provide this feature, a Neural
Machine Learning translates the text into our familiar language
How machine learning algorithms work?
 A machine learning algorithm works by learning patterns and
relationships from data to make predictions or decisions
without being explicitly programmed for each task.
Machine Learning Lifecycle
Machine Learning Lifecycle
 Defining the Problem: Clearly identify the real-world problem to be solved.
 Data Collection: Gather necessary data from various sources.
 Data Cleaning and Preprocessing: Resolve data quality issues and prepare the data for
analysis.
 Exploratory Data Analysis (EDA): Analyze data to identify patterns, outliers, and trends.
 Feature Engineering and Selection: Enhance data features and select relevant ones to
improve model performance.
 Model Selection: Choose suitable models based on the problem type and data
characteristics.
 Model Training: Train the model using a split of training and validation datasets.
 Model Evaluation and Tuning: Assess and optimize the model using relevant metrics.
 Model Deployment: Implement the model in a production environment for real-time
predictions.
 Model Monitoring and Maintenance: Regularly check and update the model to maintain
accuracy.
Types of Machine Learning Algorithms
 As new data is fed to these algorithms, they learn and
optimize their operations to improve performance, developing
'intelligence' over time.
Supervised Machine Learning
 Defined as when a model gets trained on a “Labelled
Dataset”.
 Labelled datasets have both input and output parameters.
 E.G. supervised learning for image classification
 Consider a scenario where you have to build an image classifier to
differentiate between cats and dogs.
 If you feed the datasets of dogs and cats labelled images to the
algorithm, the machine will learn to classify between a dog or a cat
from these labeled images.
 When we input new dog or cat images that it has never seen before,
it will use the learned algorithms and predict whether it is a dog or a
cat.
Advantages and Disadvantages
 Advantages
 Supervised Learning models can have high accuracy as they are
trained on labelled data.
 The process of decision-making in supervised learning models is
often interpretable.
 It can often be used in pre-trained models which saves time and
resources when developing new models from scratch.
 Disadvantages of Supervised Machine Learning
 It has limitations in knowing patterns and may struggle with unseen
or unexpected patterns that are not present in the training data.
 It can be time-consuming and costly as it relies on labeled data only.
Unsupervised Machine Learning
 Algorithm discovers patterns and relationships using unlabeled
data.
 The primary goal is to discover hidden patterns, similarities, or
clusters within the data, which can be used for various
purposes, such as data exploration, visualization,
dimensionality reduction, and more.
 E.g.
 Consider that you have a dataset that contains information about the
purchases you made from the shop.
 Through clustering, the algorithm can group the same purchasing
behavior among you and other customers, which reveals potential
customers without predefined labels.
 This type of information can help businesses get target customers as
well as identify outliers
Advantages and Disadvantages
 Advantages
 Helps to discover hidden patterns and various relationships between the
data.
 Used for tasks such as customer segmentation, anomaly
detection, and data exploration.
 It does not require labeled data and reduces the effort of data labeling.
 Disadvantages
 Without using labels, it may be difficult to predict the quality of the
model’s output.
 Cluster Interpretability may not be clear and may not have meaningful
interpretations.
 Uses techniques such as autoencoders and dimensionality reduction that
can be used to extract meaningful features from raw data.
Reinforcement Machine Learning
 A learning method that interacts with the environment by
producing actions and discovering errors.
 Trial, error, and delay are the most relevant characteristics
of reinforcement learning.
 Model keeps on increasing its performance using Reward
Feedback to learn the behavior or pattern.
 These algorithms are specific to a particular problem e.g. Google Self
Driving car, AlphaGo where a bot competes with humans and even
itself to get better and better performers in Go Game. Each time we
feed in data, they learn and add the data to their knowledge which is
training data. So, the more it learns the better it gets trained and
hence experienced.
Example
 Consider that you are training an AI agent to play a game like chess.
The agent explores different moves and receives positive or negative
feedback based on the outcome. Reinforcement Learning also finds
applications in which they learn to perform tasks by interacting with
their surroundings.
Advantages and Disadvantages
 Advantages
 It has autonomous decision-making that is well-suited for tasks and that
can learn to make a sequence of decisions, like robotics and game-playing.
 This technique is preferred to achieve long-term results that are very
difficult to achieve.
 It is used to solve a complex problems that cannot be solved by
conventional techniques.
 Disadvantages
 Training Reinforcement Learning agents can be computationally expensive
and time-consuming.
 Reinforcement learning is not preferable to solving simple problems.
 It needs a lot of data and a lot of computation, which makes it impractical
and costly.
Semi-Supervised Learning:
Supervised + Unsupervised Learning
 Works between the supervised and unsupervised learning it
uses both labelled and unlabelled data.
 It’s particularly useful when obtaining labeled data is costly,
time-consuming, or resource-intensive.
 Semi-supervised learning is chosen when labeled data requires
skills and relevant resources in order to train or learn from it.
 We use these techniques when we are dealing with data that is a
little bit labeled and the rest large portion of it is unlabeled. We can
use the unsupervised techniques to predict labels and then feed
these labels to supervised techniques. This technique is mostly
applicable in the case of image data sets where usually all images
are not labeled
Example
 Consider that we are building a language translation model,
having labeled translations for every sentence pair can be
resources intensive. It allows the models to learn from labeled
and unlabeled sentence pairs, making them more accurate.
This technique has led to significant improvements in the
quality of machine translation services.
Advantages and Disadvantages
 Advantages
 It leads to better generalization as compared to supervised learning,
as it takes both labeled and unlabeled data.
 Can be applied to a wide range of data.
 Disadvantages
 More complex to implement compared to other approaches.
 It still requires some labeled data that might not always be available
or easy to obtain.
 The unlabeled data can impact the model performance accordingly.
Current Trends
 Integrating machine learning into the IoT opens up new paths
for data analytics and real-time decision-making.
 Enhanced Predictive Analytics: The combination of IoT data and
machine learning algorithms enables more accurate predictive
models, improving decision-making in real time.
 Smart Automation: IoT devices equipped with machine learning
capabilities can autonomously adjust operations based on data
analysis, leading to more efficient processes and reduced downtime.