Page 117 - Ai_C10_Flipbook
P. 117
• Fraud and Risk Detection: The Banking and finance sector uses machine learning to analyse customer data, like
profiling, past expenditures, and other key variables to assess risks, defaults, and potential failures. This helps
them minimise bad debts and losses by offering banking products based on a customer’s purchasing power.
• Object Classification: ML model learns to identify and name objects in images and videos. For example, if
you show it pictures of animals like cats and dogs, it learns what each looks like. Later, when you give it a new
picture, it can tell whether it’s a cat or a dog.
DL
DL stands for Deep Learning. It is a subset of Machine Learning inspired by the structure and function of neurons
in the human brain, leading to the development of Artificial Neural Networks (ANNs). Deep Learning involves
implementing neural networks to process high-dimensional data, extract meaningful insights, and provide
solutions to complex problems.
Input Layer Hidden Layers Output Layer
In the hierarchy of Artificial Intelligence (AI), AI represents the broader field, with Machine Learning (ML) as a
subset, and Deep Learning (DL) as a further specialised subset of ML. While ML focuses on algorithms to learn
from data, DL represents the next evolution of ML. DL is capable of handling large datasets and complex tasks with
multi-layered neural networks.
Difference between Deep Learning and Machine Learning is shown below:
Parameters Machine Learning Deep Learning
Data Machine Learning algorithm can easily work with When the size of the data is small,
Dependency smaller data set. a Deep Learning algorithm does
not perform well as a deep learning
algorithm needs large amounts of data
to understand perfectly.
Hardware Machine Learning algorithms can work on low end
Deep Learning algorithms are heavily
Dependency machines as well.
dependent on high-end machines.
Problem When we are solving a problem using a traditional Deep Learning algorithm solves the
Solving machine learning algorithm it is generally recommended problem end to end.
Approach that we first break down the problem into different
sub parts and solve them individually and then finally
combine them to get the desired result.
Advanced Concepts of Modeling in AI 115

