Please enable JavaScript.
Coggle requires JavaScript to display documents.
Introduction to Artificial Intelligence - Coggle Diagram
Introduction to Artificial Intelligence
State of the art
AI has advanced in areas like game playing, robotics, speech recognition, and machine translation.
Game Playing: AI has revolutionized gaming through advanced strategies, learning, and simulation.
Reinforcement Learning in Games
AlphaGo, AlphaZero, MuZero
Self-play and reward optimization
Policy gradient methods
Multi-Agent Collaboration and Competition
OpenAI Five in Dota 2 -
https://openai.com/index/openai-five-defeats-dota-2-world-champions/
StarCraft II agents (DeepMind)
Agent communication and strategy sharing
Simulation-Based Training
Game environments for training real-world AI
Sim2Real transfer learning
Unity ML-Agents Toolkit
Procedural Content Generation
AI-generated game levels
Personalized gaming experiences
AI Dungeon (text adventure games) -
https://aidungeon.com/
Robotic Vehicles (Self-Driving Cars)
: Self-driving technology combines perception, decision-making, and control.
Perception Systems
Camera, radar, and LIDAR integration
Object detection and semantic segmentation
3D environment reconstruction
Path Planning and Control
Route planning using AI
Behavior prediction of other vehicles
Motion control algorithms
Simulation and Testing
Virtual environments for training and validation
Companies like Waymo and NVIDIA using simulators - NVIDIA Simulations:
https://www.nvidia.com/en-us/solutions/design-and-simulation/
Safety and edge case testing
Speech Recognition
: Speech-to-text systems are central to voice assistants and real-time transcription.
Automatic Speech Recognition (ASR)
Whisper (OpenAI), Google ASR
End-to-end deep learning models
Handling accents and noisy environments
Real-Time Transcription and Translation
Live captioning for video calls
Speech-to-speech translation (Meta's SeamlessM4T)
Subtitling tools for broadcasters
Autonomous Planning and Scheduling
: AI handles complex decision-making and dynamic resource management.
Task and Motion Planning
Robotics path planning
Search-based and learning-based approaches
Use in warehouses and space missions
Resource Allocation and Optimization
Managing limited resources in real-time
Airline and crew scheduling
Grid energy scheduling
Multi-Agent Coordination
Drone swarms and delivery bots
Decentralized decision-making
Game-theoretic planning
Spam Fighting
AI secures communication channels by detecting and blocking spam and malicious content.
Email Spam Detection
Bayesian filtering
Transformer-based filtering
Phishing and malware detection
Social Media Moderation
Detecting spammy comments and bots
Hate speech and misinformation filtering
Logistics Planning
AI streamlines supply chains, delivery, and warehouse operations.
Route Optimization
Last-mile delivery algorithms
Dynamic rerouting with real-time traffic data
Inventory Management
Predictive restocking
AI for demand forecasting
Fleet Management
AI for vehicle health prediction
Load balancing across vehicles
Robotics
: AI enables robots to perceive, decide, and act autonomously in various environments.
Computer Vision in Robotics
Object recognition and tracking
Depth estimation and SLAM (Simultaneous Localization and Mapping)
Reinforcement Learning for Control
Learning complex tasks like walking, grasping
Examples: Boston Dynamics robots
Human-Robot Interaction (HRI)
Emotion recognition and social cues
Assistive robots in healthcare and homes
Edge AI and Real-Time Processing
Low-latency decision-making on-device
AI chips for robotics (e.g., NVIDIA Jetson)
Machine Translation
: Machine translation has become more fluent and culturally aware.
Neural Machine Translation (NMT)
Transformer models (e.g., T5, mBART, MarianMT)
Contextual translation vs. phrase-based
Zero-Shot and Few-Shot Translation
Translating between languages without direct training data
Meta's No Language Left Behind (NLLB)
Multimodal Translation
Combining text, image, and speech for better context
Applications in subtitles and AR/VR
Disciplines Contributing to AI
Economics: Game theory.
Mathematics: Logic, probability, algorithms.
Philosophy: Knowledge and action.
Neuroscience: How the brain works.
Psychology: Information processing models.
Computer Engineering: Building intelligent systems.
Control Theory: Stable feedback systems.
Linguistics: Natural language processing (NLP).
Definitions of AI
AI is about building intelligent entities that can think, reason, and act like humans or rationally.
The goal is to create systems that can perform tasks requiring human-like intelligence, such as understanding language, recognizing patterns, and solving problems.
Types of AI
Types of AI
https://lumenalta.com/insights/what-are-the-different-types-of-ai
Super AI:
Smarter than humans (just an idea).
General AI:
Human-like (not real yet).
Narrow AI:
Task-specific (e.g., Chess games, Siri, spam filters).
Types of AI Behavior and Reasoning
Act Rational:
Create autonomous agents that perceive, adapt, and pursue goals.
Think Rational:
Use logic and reasoning (e.g., Aristotle’s principles).
Think Human:
Study how humans think (e.g., cognitive science, brain imaging).
Act Human:
Mimic human behavior (e.g., NLP, robotics, machine learning).
Key AI Concepts
Machine Learning (ML)
Learning from data without explicit programming.
Identifies patterns and makes predictions.
Example: Spam filters in emails.
Deep Learning
Subset of ML with neural networks.
Processes complex data like images, text, and speech.
Example: Facial recognition systems.
Natural Language Processing (NLP)
Enables computers to understand and generate human language.
Tasks include translation, sentiment analysis, and summarization.
Example: Google Translate, chatbots.
Computer Vision
Allows computers to interpret visual information (images and videos).
Used in object detection and image classification.
Example: Self-driving cars detecting obstacles.
Robotics & Automation
Robots performing tasks autonomously using AI.
Automation for repetitive or complex tasks in industries.
Example: Industrial robots, drones for delivery.
History of AI
Alan Turing and the turing test
Alan Turing: Founder of computer science, Mthamatician, philosoper, codebreaker
The turing test
Alan Turing proposed the Turing Test to determine if a machine can exhibit intelligent behavior indistinguishable from a human. If a machine can deceive a human into thinking it’s human, it passes the test.
Strong AI vs. Weak AI
Strong AI: Machines can think and have consciousness.
Weak AI: Machines can simulate intelligence but don’t have consciousness.
The Turing Test:
https://www.techtarget.com/searchenterpriseai/definition/Turing-test
The Chinese Room Argument
John Searle (1980): A person in a room follows instructions to respond to Chinese messages but doesn’t understand Chinese.
Definition of The Chinese Room Argument
A philosophical thought experiment by John Searle that argues a computer running a program cannot be said to "understand" or have "consciousness," even if it appears to understand language. It demonstrates that mere symbol manipulation (syntax) is not sufficient for true understanding (semantics).
Chinese Room Argument
https://plato.stanford.edu/entries/chinese-room/
Consciousness
Involves self-awareness, perception, judgment, and understanding the relationship between self and environment.
subjectivity - have a perspective
self awareness
sentience - perception
sapience - act with judgement
perceive relationship between self and environment