Assignment 3

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Assignment 3

Supervised By

Dr Mohamed Abo Rizka

Prepared By

Saif Allah Mohamed Bakry


What is Gradient descent method?
Gradient Descent is a machine learning iterative optimization algorithm that is
frequently used to train machine learning and deep learning models. It assists
in determining a function's local minimum

Types of Gradient Descent


1. Batch Gradient Descent
2. Stochastic gradient descent
3. Minibatch Gradient Descent

Batch Gradient Descent:


After evaluating all training examples, batch gradient descent (BGD) is implemented to
find the error for each point in the training set and update the model. This is referred to as
the training epoch. Simply put, it is a greedy approach in which we must add over all
examples for each update

Advantages of Batch gradient descent:


It generates less noise in comparison to other gradient descent
It produces stable gradient descent convergence
Because all resources are utilized for all training samples, it is efficient and robust.
___________________________________________________________________________________________

Stochastic gradient Descent


Stochastic gradient descent (SGD) is a gradient descent algorithm that runs one training example
per iteration. In other words, it processes a training epoch for each example in a dataset and
updates the parameters of each training example one at a time. It is easier to store in allocated
memory because it only requires one training example at a time

Page | 2
Advantages of Stochastic gradient descent:
It is simplest way to allocate in the desired memory.
It is quicker to compute than batch gradient descent.
For large datasets, it is more proficient.
___________________________________________________________________________________________

MiniBatch Gradient Descent:


Mini Batch gradient descent integrates batch gradient descent with stochastic gradient
descent. It divides the training datasets into small batches and then updates each batch
separately. Splitting training datasets into smaller batches strikes a balance between batch
gradient descent's computational efficiency and the speed of stochastic gradient descent
Advantages of Mini Batch gradient descent:
It is more easy to fit in allocated memory
It is computationally efficient
It produces stable gradient descent convergence
___________________________________________________________________________________________

What is weights:
Weights are real values that are associated with each input/feature and convey the
significance of that correlating feature in predicting the final output

What is Bias:
The constant added to the product of features and weights is known as bias. It is used to cancel
out the outcome. It aids models in shifting the activation function to the positive or negative side

What is activation function:


An activation function is a function in an artificial neuron that generates an output based
on the inputs. Activation functions in artificial neurons are an essential part of the role that
artificial neurons perform in modern artificial neural networks.

Page | 3
What is Geo AI?
is a branch of study that integrates the use of geographic information systems (GIS) and
artificial intelligence (AI) to analyze spatial data and make forecasts or decisions based on
it. It is an interdisciplinary field that combines geographers', computer scientists', and data
scientists' expertise to create innovative solutions to real-world problems

What is AI GIS
AI GIS is a mixture of AI technology and several GIS functions, including spatial data
processing and analysis algorithms (GeoAI) that combine AI technology, and a general
term for a series of AI and GIS consensual empowerment technologies. In recent years, AI
GIS has gradually become the core focus of geoscience both research and application

What is GIS AI
When confronted with AI recognition findings, GIS for AI empowers AI by utilizing its
geographical visualization and spatial analytic capabilities to further process and mine data.
Map visualization applications such as traffic flow monitoring and city component
management can provide decision makers with a more intuitive form of information
expression, while in-depth processing and mining of AI extraction results can enable real-
time geo fence alerts, vehicle tracking, and other applications.

Page | 4

You might also like