Artificial Intelligence - UNIT - 1 Topic : 3. Foundation of AI and History of AI
UNIT - 1
3. FOUNDATIONS OF AI AND HISTORY OF AI
Part A: Foundations of
Artificial Intelligence
1. Introduction
The foundation of AI lies in several disciplines that provide the
necessary concepts, theories, tools, and methodologies for developing
intelligent systems. These disciplines work together to help AI systems think,
reason, learn, and make decisions.
2. Major Foundations of AI
Artificial
Intelligence is a multidisciplinary field, meaning it draws ideas and
techniques from many areas of study. Here’s how each field contributes:
a. Mathematics
Mathematics forms the core framework behind most AI models and
algorithms.
- Logic (Propositional
& Predicate Logic):
Helps AI make decisions through rules and reasoning.
- Probability &
Statistics: Essential for
handling uncertainty and making predictions (e.g., Bayesian
Networks).
- Linear Algebra: Used in neural networks, image
processing, and data transformations.
- Calculus: Especially derivatives and gradients,
used in training AI models through optimization methods like
gradient descent.
- Graph Theory: Supports search algorithms, planning, and
game strategies (e.g., pathfinding).
Why it matters: Without math, AI cannot
learn, reason, or optimize decisions.
b. Computer Science and Engineering
This foundation gives AI the tools and platforms to operate efficiently.
- Algorithms & Data
Structures: Enable fast and
effective problem-solving and data handling.
- Programming Languages: Languages like Python, LISP, Prolog,
and Java are used to develop AI systems.
- Computational
Complexity: Helps understand the time
and space AI algorithms will use.
- Hardware Platforms: AI relies on GPUs, TPUs, and
specialized chips for high-speed processing, especially in deep
learning and robotics.
Why it matters: AI needs strong
computational power and logic to function practically.
c. Psychology
AI tries to mimic human intelligence, so understanding human
mental processes is key.
- Focuses on how humans perceive,
learn, remember, solve problems, and feel emotions.
- Helps in building human-like
AI models (e.g., cognitive architectures).
Why it matters: To create AI that
"thinks" like humans, we must first understand how humans think.
d. Philosophy
Philosophy explores the fundamental questions of intelligence,
consciousness, and morality.
- Involves concepts like
consciousness, ethics, free will, and rational reasoning.
- Helps define what
it means to “think” or be “intelligent.”
- Theories like Dualism
(mind-body), Rationalism (reason over experience) influence how AI
reasoning is structured.
Why it matters: Guides the ethical and
conceptual boundaries of AI development.
e. Linguistics
Language is a big part of human intelligence, and linguistics helps AI
understand and generate it.
- Supports Natural
Language Processing (NLP) for reading, writing, and understanding
human language.
- Enables speech
recognition, text analysis, and translation.
- Helps AI understand grammar,
meaning (semantics), and context.
Why it matters: AI must understand human
language to interact with us naturally.
f. Neuroscience
AI takes inspiration from how the human brain processes information.
- Helps in designing artificial
neural networks by mimicking brain structures.
- Aids in developing brain-computer
interfaces.
- Supports building
models based on cognitive functions.
Why it matters: Neuroscience shows how natural
intelligence works, inspiring better artificial versions.
🔁 In Summary:
|
Foundation |
AI Contribution |
|
Mathematics |
Provides the structure for learning, logic, and prediction |
|
Computer Science |
Offers tools, languages, algorithms, and hardware |
|
Psychology |
Helps model human-like behavior and learning |
|
Philosophy |
Addresses reasoning, ethics, and the meaning of intelligence |
|
Linguistics |
Enables language understanding and communication |
|
Neuroscience |
Inspires neural networks and cognitive modeling |
3. Interdisciplinary Nature
of AI
AI draws strength from the integration of these areas:
|
Discipline |
Contribution to AI |
|
Mathematics |
Logical representation, search, optimization |
|
CS/Engineering |
Algorithms, computing power, software |
|
Psychology |
Understanding cognition and decision-making |
|
Philosophy |
Logical reasoning and ethics |
|
Linguistics |
Natural language understanding |
|
Neuroscience |
Brain-like learning models |
4. Summary: Foundations of
AI
- AI
is not limited to computer science.
- Combines
ideas from many fields to simulate intelligence.
- Understanding
these foundations helps build robust AI systems.
1. Early Beginnings (Before
1950)
- 1943 –
McCulloch & Pitts propose a simplified model of the neuron.
- 1949 –
Hebb’s rule: "Neurons that fire together wire together."
- 1950 –
Alan Turing publishes “Computing Machinery and Intelligence” and
proposes the Turing Test.
2. Birth of AI (1956 – 1970)
- 1956 –
The term "Artificial Intelligence" coined at the Dartmouth
Conference by John McCarthy.
- Early
projects focused on:
- Solving algebra problems
- Playing games (e.g., checkers, chess)
- The Logic Theorist – first AI program by Newell & Simon
3. Golden Years (1970 –
1980)
- Development
of LISP language (dominant AI language for decades).
- Creation
of Expert Systems:
- MYCIN –
for medical diagnosis
- DENDRAL –
for chemical analysis
- AI
research received significant government and industrial funding.
4. First AI Winter (1980s)
- Overhyped expectations failed to deliver practical results.
- Funding
cuts led to AI Winter (reduced interest and support).
5. Revival and Second Wave
(1990 – 2010)
- Introduction
of Machine Learning and Neural Networks.
- Growth
in computing power and data availability.
- Major
events:
- 1997 –
IBM’s Deep Blue defeats Garry Kasparov in chess.
- 2005 –
Stanford’s robot wins DARPA Grand Challenge.
6. Modern Era (2010 –
Present)
- Rise
of Big Data and Cloud Computing
- Development
of Deep Learning and Transformers
- Milestones:
- 2011 –
IBM Watson wins Jeopardy!
- 2016 –
Google DeepMind’s AlphaGo defeats world champion
- 2020s –
OpenAI's GPT, ChatGPT, DALL·E, etc.
7. Timeline Overview
|
Year |
Milestone |
|
1950 |
Alan Turing proposes the Turing Test |
|
1956 |
Dartmouth Conference – AI officially born |
|
1980 |
Expert Systems boom (e.g., MYCIN, DENDRAL) |
|
1997 |
Deep Blue beats world chess champion |
|
2011 |
IBM Watson wins Jeopardy! |
|
2016 |
AlphaGo defeats world Go champion |
|
2022 |
ChatGPT launched (language generation AI) |
8. Summary: History of AI
- AI
has evolved from symbolic reasoning to deep learning.
- Experienced
highs and lows (AI Winters).
- Currently
in a phase of rapid innovation and real-world applications.
Comments
Post a Comment