Deep dive into AI
Deep Dive into AI
Presenter: Piyush Kumar Arya
AI Learning Methods
AI can learn in different ways based on data and feedback. Three major learning methods are:
Supervised Learning:
Learns from labeled data.
Unsupervised Learning:
Finds patterns in unlabeled data.
Reinforcement Learning:
Learns by interacting with an environment.
Supervised vs Unsupervised Learning
Supervised Learning
Uses labeled data
Learns from correct answers
Common Algorithms:
Decision Trees
Support Vector Machines (SVM)
Neural Networks
Examples:
Spam detection (emails labeled spam or not spam)
Image classification (cat vs. dog photos)
Unsupervised Learning
Uses unlabeled data
Finds patterns on its own
Common Algorithms:
Clustering (K-Means)
Principal Component Analysis (PCA)
Autoencoders
Examples:
Customer segmentation in marketing
Anomaly detection in fraud detection
Reinforcement Learning (RL)
AI learns by trial and error, improving based on rewards and punishments.
Key Components:
Agent: The AI making decisions.
Environment: Where the agent interacts.
Reward System: Encourages good actions, penalizes bad ones.
Examples:
Self-driving cars adjusting speed and turns
AlphaGo (AI beating human Go players)
Robotics (robots learning to walk)
Popular Algorithms:
Q-Learning
Deep Q Networks (DQN)
Policy Gradients
Reinforcement Learning Process
Observe: The agent observes the environment.
e.g., "Ouch! -50 points"
Select action: The agent chooses an action using a policy.
Action! The agent takes an action.
Get reward or penalty: The agent receives feedback (reward or penalty).
e.g., "bad! Next time avoid it."
Update policy: A learning step where the agent updates its policy based on feedback.
Iterate until an optimal policy is found.
Deep Learning & Neural Networks
Deep Learning is a subset of Machine Learning using artificial neural networks inspired by the human brain.
Neural Network Structure:
Input Layer: Takes in data.
Hidden Layers: Processes and learns patterns.
Output Layer: Provides final decision/prediction.
Applications and Popular Architectures
Computer Vision:
Image recognition
Facial recognition
Natural Language Processing (NLP):
Chatbots
Translation (Google Translate)
Self-Driving Cars:
Object detection and decision-making
Popular Architectures:
Convolutional Neural Networks (CNNs) for images
Recurrent Neural Networks (RNNs) for sequential data
Transformers for NLP