AI Notes
What is AI ?
Artificial Intelligence (AI) is a field of computer science focused on developing
systems that can perform tasks typically requiring human intelligence. These
tasks include learning from experience, reasoning, problem-solving,
understanding natural language, recognizing images, and making decisions.
While AI seeks to simulate or replicate aspects of human cognition, its methods
often differ from human thought processes.
Key Branches of AI include:
<u>Machine Learning (ML):</u> A subset of AI where systems learn from
data without explicit programming.
<u>Natural Language Processing (NLP)</u>: Enabling machines to
understand and generate human language.
<u>Computer Vision:</u> Allowing machines to interpret visual information.
<u>Robotics:</u> Combining AI with physical systems for automation.
History of Artificial Intelligence
AI’s history spans seven decades, with key milestones:
1940s: Turing’s Turing Machine and Test; McCulloch-Pitts’ neural networks.
1950s: McCarthy coins “AI” at Dartmouth (1956); Logic Theorist emerges.
1960s: ELIZA and General Problem Solver; Cold War fuels funding.
1970s: AI winter due to overhype and limited computing power.
1980s: Expert systems (e.g., MYCIN); Japan’s Fifth Generation Project.
1990s: Machine learning rises; Deep Blue beats Kasparov (1997).
2000s: Big data, GPUs, and Deep Learning (e.g., AlexNet, 2012).
2010s: NLP (BERT), computer vision, and AlphaGo (2016); AI goes
mainstream.
2020s: Generative AI (ChatGPT, DALL-E); focus on ethics and regulation.
AI Notes 1
, AI vs Human Intelligence
Aspect Artificial Intelligence (AI) Human Intelligence
Biological, shaped by
Engineered, algorithm-based,
Nature evolution, experience,
operates within set parameters
emotions
Combines reasoning,
Deterministic/probabilistic; excels in
Processing intuition, emotional
pattern recognition
processing
Data-driven, requires explicit training, Dynamic, generalizes from
Learning
limited adaptability limited data, highly adaptive
Fast, handles vast data for repetitive Slower, excels in nuanced,
Speed & Scale
tasks context-dependent tasks
Driven by imagination,
Creativity & Simulates creativity, lacks genuine
emotions, subjective
Emotion emotions
experience
Narrow, domain-specific; general AI Broadly general, applies
Scope
theoretical knowledge across domains
No inherent morality, relies on Guided by values, norms,
Ethics & Morality
programmed rules subjective ethical reasoning
Low energy (~20 watts),
Energy & High computational energy, no
needs sleep, food, well-
Maintenance biological needs
being
Components of AI
1. Learning
Learning is at the heart of AI, allowing systems to improve over time. AI can
learn in three main ways:
Supervised Learning: Involves training AI with labeled data, where the
system is taught by example and can predict outcomes.
Unsupervised Learning: The AI is given data without explicit instructions
and discovers patterns or relationships on its own.
Reinforcement Learning: This involves learning through trial and error,
where the AI receives feedback on its actions and adjusts accordingly.
2. Reasoning and Decision Making
AI Notes 2
, AI systems use reasoning to analyze data and make decisions. They can draw
conclusions based on pre-programmed rules, probabilities, or models learned
from previous experiences. This allows AI to make informed decisions, similar
to human reasoning processes.
3. Problem Solving
AI can identify problems and come up with solutions using various strategies.
Whether it’s playing a game or diagnosing a technical issue, AI uses algorithms
to break down complex problems and adapt to new situations.
4. Perception
Perception in AI refers to how systems collect and interpret data from the
world. Using sensors like cameras and microphones, AI can see, hear, and
understand its surroundings. This component is vital for tasks such as
recognizing images or interpreting speech.
5. Processing Language
Language processing, often referred to asNatural Language Processing (NLP),
is how AI understands and interacts with human language. NLP enables AI to
comprehend, generate, and respond to text or speech, making it a crucial part
of AI-powered assistants like Siri or chatbots.
Types of AIs
Artificial Intelligence (AI) can be classified based on its capabilities and
functionality. Below is a concise explanation of the main types of AI, organized
into two primary frameworks: capability-based and functionality-based.
1. Capability-Based Types of AI
This classification focuses on the level of intelligence and autonomy AI systems
exhibit.
Narrow AI (Weak AI)
Definition: AI designed for specific tasks, operating within predefined
boundaries. It excels in narrow domains but lacks general intelligence.
Characteristics: Task-specific, rule-based or data-driven, no self-
awareness or broad reasoning.
AI Notes 3
, Examples:
Virtual assistants (e.g., Siri, Alexa).
Image recognition systems (e.g., facial recognition).
Recommendation algorithms (e.g., Netflix, Spotify).
General AI (Strong AI)
Definition: Hypothetical AI with human-like intelligence, capable of
performing any intellectual task a human can across diverse domains.
Characteristics: General problem-solving, adaptability, and reasoning
across contexts; not yet achieved.
Examples: None exist today; remains a theoretical goal in AI research.
Superintelligent AI
Definition: Hypothetical AI surpassing human intelligence in all areas,
including creativity, problem-solving, and emotional intelligence.
Characteristics: Far exceeds human capabilities; raises ethical and
existential concerns.
Examples: Purely speculative (e.g., sci-fi depictions like Skynet in
Terminator).
2. Functionality-Based Types of AI
This classification focuses on the operational mechanisms and techniques AI
systems use.
Reactive AI
Definition: AI that responds to specific inputs without memory or
learning, focusing solely on the current state.
Characteristics: No past or future context; simple, rule-based
responses.
Examples: IBM’s Deep Blue (chess AI), basic chatbots with fixed
responses.
<u>Limited Memory AI</u>
Definition: AI that uses historical data to inform decisions, enabling
short-term memory for context.
AI Notes 4
What is AI ?
Artificial Intelligence (AI) is a field of computer science focused on developing
systems that can perform tasks typically requiring human intelligence. These
tasks include learning from experience, reasoning, problem-solving,
understanding natural language, recognizing images, and making decisions.
While AI seeks to simulate or replicate aspects of human cognition, its methods
often differ from human thought processes.
Key Branches of AI include:
<u>Machine Learning (ML):</u> A subset of AI where systems learn from
data without explicit programming.
<u>Natural Language Processing (NLP)</u>: Enabling machines to
understand and generate human language.
<u>Computer Vision:</u> Allowing machines to interpret visual information.
<u>Robotics:</u> Combining AI with physical systems for automation.
History of Artificial Intelligence
AI’s history spans seven decades, with key milestones:
1940s: Turing’s Turing Machine and Test; McCulloch-Pitts’ neural networks.
1950s: McCarthy coins “AI” at Dartmouth (1956); Logic Theorist emerges.
1960s: ELIZA and General Problem Solver; Cold War fuels funding.
1970s: AI winter due to overhype and limited computing power.
1980s: Expert systems (e.g., MYCIN); Japan’s Fifth Generation Project.
1990s: Machine learning rises; Deep Blue beats Kasparov (1997).
2000s: Big data, GPUs, and Deep Learning (e.g., AlexNet, 2012).
2010s: NLP (BERT), computer vision, and AlphaGo (2016); AI goes
mainstream.
2020s: Generative AI (ChatGPT, DALL-E); focus on ethics and regulation.
AI Notes 1
, AI vs Human Intelligence
Aspect Artificial Intelligence (AI) Human Intelligence
Biological, shaped by
Engineered, algorithm-based,
Nature evolution, experience,
operates within set parameters
emotions
Combines reasoning,
Deterministic/probabilistic; excels in
Processing intuition, emotional
pattern recognition
processing
Data-driven, requires explicit training, Dynamic, generalizes from
Learning
limited adaptability limited data, highly adaptive
Fast, handles vast data for repetitive Slower, excels in nuanced,
Speed & Scale
tasks context-dependent tasks
Driven by imagination,
Creativity & Simulates creativity, lacks genuine
emotions, subjective
Emotion emotions
experience
Narrow, domain-specific; general AI Broadly general, applies
Scope
theoretical knowledge across domains
No inherent morality, relies on Guided by values, norms,
Ethics & Morality
programmed rules subjective ethical reasoning
Low energy (~20 watts),
Energy & High computational energy, no
needs sleep, food, well-
Maintenance biological needs
being
Components of AI
1. Learning
Learning is at the heart of AI, allowing systems to improve over time. AI can
learn in three main ways:
Supervised Learning: Involves training AI with labeled data, where the
system is taught by example and can predict outcomes.
Unsupervised Learning: The AI is given data without explicit instructions
and discovers patterns or relationships on its own.
Reinforcement Learning: This involves learning through trial and error,
where the AI receives feedback on its actions and adjusts accordingly.
2. Reasoning and Decision Making
AI Notes 2
, AI systems use reasoning to analyze data and make decisions. They can draw
conclusions based on pre-programmed rules, probabilities, or models learned
from previous experiences. This allows AI to make informed decisions, similar
to human reasoning processes.
3. Problem Solving
AI can identify problems and come up with solutions using various strategies.
Whether it’s playing a game or diagnosing a technical issue, AI uses algorithms
to break down complex problems and adapt to new situations.
4. Perception
Perception in AI refers to how systems collect and interpret data from the
world. Using sensors like cameras and microphones, AI can see, hear, and
understand its surroundings. This component is vital for tasks such as
recognizing images or interpreting speech.
5. Processing Language
Language processing, often referred to asNatural Language Processing (NLP),
is how AI understands and interacts with human language. NLP enables AI to
comprehend, generate, and respond to text or speech, making it a crucial part
of AI-powered assistants like Siri or chatbots.
Types of AIs
Artificial Intelligence (AI) can be classified based on its capabilities and
functionality. Below is a concise explanation of the main types of AI, organized
into two primary frameworks: capability-based and functionality-based.
1. Capability-Based Types of AI
This classification focuses on the level of intelligence and autonomy AI systems
exhibit.
Narrow AI (Weak AI)
Definition: AI designed for specific tasks, operating within predefined
boundaries. It excels in narrow domains but lacks general intelligence.
Characteristics: Task-specific, rule-based or data-driven, no self-
awareness or broad reasoning.
AI Notes 3
, Examples:
Virtual assistants (e.g., Siri, Alexa).
Image recognition systems (e.g., facial recognition).
Recommendation algorithms (e.g., Netflix, Spotify).
General AI (Strong AI)
Definition: Hypothetical AI with human-like intelligence, capable of
performing any intellectual task a human can across diverse domains.
Characteristics: General problem-solving, adaptability, and reasoning
across contexts; not yet achieved.
Examples: None exist today; remains a theoretical goal in AI research.
Superintelligent AI
Definition: Hypothetical AI surpassing human intelligence in all areas,
including creativity, problem-solving, and emotional intelligence.
Characteristics: Far exceeds human capabilities; raises ethical and
existential concerns.
Examples: Purely speculative (e.g., sci-fi depictions like Skynet in
Terminator).
2. Functionality-Based Types of AI
This classification focuses on the operational mechanisms and techniques AI
systems use.
Reactive AI
Definition: AI that responds to specific inputs without memory or
learning, focusing solely on the current state.
Characteristics: No past or future context; simple, rule-based
responses.
Examples: IBM’s Deep Blue (chess AI), basic chatbots with fixed
responses.
<u>Limited Memory AI</u>
Definition: AI that uses historical data to inform decisions, enabling
short-term memory for context.
AI Notes 4
