Scribd
Upload
14 views

Deep Learning NLP and Computer Vision

Uploaded by

istiaq8888
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
14 views

Deep Learning NLP and Computer Vision

Uploaded by

istiaq8888
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

aiQuest Intelligence & Study Mart

Web: https://www.aiquest.org

Email: [email protected]

Cell: +8801704265972

Deep Learning, NLP & Computer Vision

Course Type: Live Class (Night 8pm-10pm)


Total Hours: 50+
Total Class: 22
Every Week: 2 Class
Fees: 5000 Taka

Course Instructor:
Mr. Mejbah Ahammad
Deep Learning Instructor at aiQuest Intelligence
Tableau Certified Data Scientist
Class: 01

1. Why AI?
2. Introduction to Deep Learning
3. Use cases for deep learning
4. Deep Learning's importance
5. How do neural networks work?
6. The Function of Neural Networks

Class: 02

Software Installation:
1. Latest Python
2. PyCharm
3. Visual Studio
4. Anaconda
5. Atom

Python Basics:
1. Introduction of Python
2. Python objects and Data structure basics
3. Python statements
4. Variables and names
5. Strings and text
6. Reading and writing files
7. Classes and objects
8. Functions
9. File handling
10. Conditionals and control flow
11. List and dictionaries Loops

Class: 03
1. Neuron
2. Weights
3. Bias
4. CNN (Convolutional neural network)
5. Pooling
6. Data Augmentation
o Position augmentation
o Color augmentation
7. Activation Functions
8. Neural Network
9. MLP (Multi Layer perceptron)
10. Forward Propagation
11. Cost Function
12. Padding
13. Gradient Descent
14. Learning Rate
15. Backpropagation
16. Batches
17. Epochs
18. Dropout
19. Batch Normalization
20. Filters
21. CNN (Convolutional neural network)
22. Pooling
23. Padding
24. Data Augmentation
25. Recurrent Neural Network
26. Gradient Problem
o Vanishing Gradient Problem
o Exploding Gradient Problem

Class: 04 Implementation 01
1. Root Implementation of Neural Network Using Pytorch
a. Layers
b. Activation functions
c. Loss function
d. Optimizer
e. Weights and biases
f. Layers
g. Activation functions

Class: 05: Image Classification with Convolutional Neural Network.


Project: 01
1. Visualizations
2. Dealing with Class imbalance
3. Fill missing values (labels, features and, etc.)
4. Normalization
5. Pre-processing of the images
6. Model training
7. Compiling the model
8. Fitting the model
9. Model evaluation

Class: 06 Making Good Understanding With the following Neural


Architecture

1. LeNet
2. AlexNet
3. ResNet
4. GoogleNet/InceptionNet
5. MobileNetV1
6. Inception v2
7. Inception v3 V4 and Inception
8. ResNet.
9. DenseNet
10. Xception

Class: 07 Medical Image Classification


Project 02:
1. Image Acquisitions
2. Image Pre-processing
3. Feature Extractions
a. Texture
b. Wavelet
c. Region Based
d. Histogram
4. Evaluation
5. Classification
Class: 08 Recurrent Neural Network

Background Theory of Recurrent Neural Network


1. Binary
2. Linear
3. Continuous
4. Nonlinear

5. Additive STM equation


6. Shunting STM equation
7. Generalized STM equation
8. MTM: Habituative Transmitter Gates and Depressing Synapses
9. LTM: Gated steepest descent learning: Not Hebbian learning

Class: 09 Basic Implementation of RNN


[Simple RNN, LSTM, BI-LSTM, etc]

1. Initialize weight
2. Forward propagation to compute predictions
3. Compute the loss
4. Back-propagation to compute gradients
5. Update weights based on gradients

Class: 10 Implementation of Basic Terminology

1. Tokenization
2. Stemming
3. Normalization
4. Lemmatization
5. Corpus
6. Bag of Words
7. Stop Words
8. n-grams
9. Parts-of-speech (POS)
10. Tagging
11. Statistical Language Modeling Similarity Measures
12. Regular Expressions
13. Sentiment Analysis
14. Syntactic Analysis
15. Semantic Analysis
16. Information Retrieval
Class: 11 Sentiment Analysis with Traditional Approach
Project 3:
1. Naïve Bayes,
2. Logistic Regression and
3. Support Vector Machines (SVM)
4. Neural Network

Class: 12 LSTM [Long Short-Term Memory] Implementation

1. Importing libraries
2. Loading data
3. Data exploration
4. Data preprocessing
5. Model building
6. Evaluating models

Class: 13 Stock Price Prediction With LSTM


Project 04:

1. Preprocessing Date and convert into Datatime format


2. Perform Exploratory Data Analysis(EDA)
3. Find the duration of dataset
4. Trend comparison between stock open price, close price, high price, low price
5. Prepare Stock Close price Make separate dataframe
6. Plot stock close price Consider only 1 year data for prediction
7. Normalizing close price
8. Split Train and Test
9. Transform Close price base on Time-series-analysis forecasting requirement
Class: 14 Text Classification with Self Attention
Project 05:

1. Gather Data.
2. Explore Your Data.
3. Choose a Model
4. Prepare Your Data.
5. Build, Train, and Evaluate Your Model.
6. Tune Hyperparameters.
7. Implementing a Simple Attention Model in Python using Keras
8. Global vs. Local Attention
9. Transformers – Attention

Class: 15 Text Summarization


Project 06:
1. Extractive Summarization
2. Abstractive Summarization
3. Reading the original text again
4. List the main points.
5. Take note of the supporting details.
6. Start with a statement that is contextual.
7. Describe the text's main idea.
8. Continue with supporting details.
9. Make a fundamental argument.

Class: 16 Disease Detection


Project: 07

1. Choice of deep learning architecture:


1. AlexNet,
2. GoogLeNet.
2. Choice of training mechanism:
1. Transfer Learning,
2. Training from Scratch.
3. Choice of dataset type:
1. Color
2. Gray scale
4. Choice of training-testing set distribution
Class: 17 Object Detection (R-CNN)
Project: 08
1. Extracting regions for each image
2. Extracting features using
3. Basic structure
4. Model architecture overview
5. Object detection works on the edge

Class: 18 Final Exam and Viva


1. Written Exam: 10%
2. Code Exam: 50%
3. MCQ Exam: 10%
4. Viva: 10%
5. Attendance: 20%

Class: 19 Capstone Project (R-CNN))

1. Find regions in the image that might contain an object.


2. These regions are called region proposals.
3. Extract CNN features from the region proposals.
4. Classify the objects using the extracted features.
5. Slow training and detection
6. Allows custom region proposal
7. Optimal run-time performance
8. Region proposal network
9. Reader and minibatch source
Class: 20 Capstone Project (Yolo)

1. Object-ness score thresholding


2. Non-maximum suppression
3. Creating the layers of the network architecture
4. Implementing the the forward pass of the network
5. Designing the input and the output pipelines

Class: 21 Career Guidelines


1. Jobs
2. CV
3. Final Projects

Join with Us:


Web: www.aiquest.org

Cell: +8801704265972

YouTube: https://www.youtube.com/StudyMart

FB Community: https://www.facebook.com/groups/StudyMart

You might also like