overview
What is Artificial Intelligence?
AI stands for "Artificial Intelligence," which is: making computers and robots think and learn like humans. Just like we learn from our experiences and use that knowledge to make decisions, AI allows machines to do the same.
Example Smartphone's voice assistant, your email filtering out spam or a self-driving car navigating the roads.
So, when we talk about AI, we're talking about giving machines the ability to learn, reason, and make decisions—just like us humans!
What are the three main subfields of AI?
Machine Learning (ML) - Focuses on machine to learn from Data Ex. Image & Speech Detection, Fraud Detection, Autonomous Vehicle etc. (Data Driven Models).
Types of ML
- Supervised - is a machine learning approach where a model is trained to predict an output based on labeled data
- Classification - is a type of learning that involves assigning data to specific categories or classes
- Regression - is a type of learning that involves predicting continuous values
- UnSupervised - is a machine learning approach where a model is trained on unlabeled data
- Clustering - Algorithms use mathematical methods to group data points based on their similarities or differences.
- Associational - Algorithms use mathematical methods to look for relationship in variables
- Dimensionality - Algorithms use mathematical methods to transform high-dimensional data into a lower-dimensional
- Reinforcement - is a machine learning approach where a model is trained by interacting with an environment
Machine Learning Algorithms
- Linear Regression
- Logistic Regression
- Decision Trees
- Random Forests
- Support Vector Machines (SVM)
Machine Learning Frameworks
- TensorFlow
- PyTorch
- Scikit-Learn
- Keras
- Apache Spark MLlib
Natural Language Processing (NLP) - Focuses on machine to interpret human language. Ex. Chat Bots, Virtual Assistants etc.
Types of layers
- Input layer
- Hidden layers
- Output layer
Deep Learning Frameworks are essential tools for building and deploying artificial neural networks
- Caffe
- Keras
- MXNet
- PyTorch
- TensorFlow
- TensorFlow 2.0+
- Theano
Convolutional Neural Network (CNN) is a deep learning algorithm specifically designed for analyzing visual data such as images and videos.
Two commonly used CNN architectures
- VGG (Visual Geometry Group).
- ResNet (Residual Network).
Fundamental operations in CNNs
- Convolutional operation: How filters are applied to input data.
- Padding: Adding additional border pixels to the input to preserve spatial dimensions.
- Stride: The number of pixels by which the filter moves after each computation.
- Feature maps: Output produced by applying filters to the input data.
- Pooling :
Robotics - Focus on designing robots to interact with physical world. Used in industries Ex. Manufacturing, Healthcare and Entertainment