Scribd
Upload
12 views

Machine Learning Presentation

This document provides an overview of machine learning workflows and processes. It discusses key steps like data cleaning, feature scaling, train-test splits, supervised vs unsupervised learning, and different machine learning models like regression, decision trees, convolutional neural networks, recurrent neural networks, generative models, and more. The goal is to explore how these processes impact predictive model performance.

Uploaded by

Bunny
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
12 views

Machine Learning Presentation

This document provides an overview of machine learning workflows and processes. It discusses key steps like data cleaning, feature scaling, train-test splits, supervised vs unsupervised learning, and different machine learning models like regression, decision trees, convolutional neural networks, recurrent neural networks, generative models, and more. The goal is to explore how these processes impact predictive model performance.

Uploaded by

Bunny
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 12

Machine Learning

Presentation
Welcome to an overview of the machine learning workflow, from data
preparation to advanced model deployment. We'll explore key processes and their
impact on predictive model performance.

by ML Team
Data Cleaning

● “In the world of machine learning, the journey from raw data to accurate predictions begins with a
critical process – Data Cleaning."

Why Data Cleaning Matters ?

● Data cleaning plays a pivotal role in enabling accurate, reliable, and insightful machine learning
models.
● Poor data quality directly impacts machine learning model performance.
● Column required or not
● Format of column
● Outliers
● Handling Missing Values
● Removing Duplicates, etc.
Feature Scaling
Min-Max Standardization Scaling Benefits
Scaling
Rescales features to a 0-1 range, Adjusts features to a mean of Essential for neural network
sensitive to outliers. zero and standard deviation of performance and consistent
one. Euclidean distances.
Train-Test Split

Split Ratio Overfitting Underfitting


Varies from 50/50 to 90/10, Model learns noise instead of Model too simple, performs
tailored to dataset size and patterns, failing on new data. poorly on training and testing
model complexity. data.
Supervised vs Unsupervised Learning
Supervised Unsupervised Learning
Learning

Trains on labeled data, ideal for prediction like Finds patterns in unlabeled data, great for customer
customer churn. segmentation.
Regression and Decision Trees

1 Regression Analysis

Quantifies relationships for forecasting and planning.

2 Decision Trees
Splits data for outcomes like credit risk assessment.
Deep Learning Model Tasks
Overview.

ANN CNN
Tackled nonlinear classification problem
Image Recognizer

RNN

RNNs help improve language translation by understanding the context of words in a sentence,
ensuring more accurate and meaningful translations in machine translation systems.
Gen AI Models
1 Food Recipe Generator Model

This model provides a recipe based on a set of ingredients or a particular type of cuisine
provided by the user, offering step-by-step directions on how to prepare a particular dish.

Image Recognizer Using Google AI


Object Classification
2
This tool uses Google's AI capabilities to identify objects or features in an image,
classifying images into thousands of categories and detecting individual objects and
faces within images.
Text to Image

Text Conditional Image Generation Model: The goal of this model is to generate
an image given a textual description. For example, given a brief input like "a
landscape with mountains and a blue sky", the model would generate an image
matching that description.

Text to Video

Text to Video Generation Model: Given a textual description, this model can
generate a short video clip. The description must be highly detailed to offer a
precise representation in the generated video.
Image to Text

Image Captioning Model: An image captioning model uses computer vision and
natural language processing to generate a textual description of an image. This
model not only detects the objects in an image but also understands the context
making it capable of generating a fitting description or caption.

Sample output:

Image to Image

Real Image to Animated Image Transformer Model: This model transforms real-
world images into animated versions. It has potential use cases in the
entertainment industry, turning real-life scenes or characters into animated
depictions.
Thank You
Dhananjay Swapneel Bharat Tyagraj Ritesh Hrutik

You might also like