Intro AI DBATU.ppt ppt useful engineering

Artificial Intelligence
 1.1 Definition of AI
 1.2 AI technique
 1.3 Criteria for success
 1.4 AI application areas
 1.5 Summary
2

1.1 Defining Artificial
Intelligence(1)
 “AI is the study of how to make computers do things which, at
the moment, people do better.” (Rich)
 “AI is the part of computer science concerned with designing
intelligent computer systems, that is, systems that exhibit
characteristics we associate with intelligent human behavior.
 understanding language, reasoning, solving problems, and so on.”
(Barr)
 “AI is the study of ideas which enable computers to do things
which make people seem intelligent.” (Winston)
 AI is the study of intelligence using the ideas and methods of
computation.” (Fahlman)
3

Defining Artificial Intelligence(2)
 “A bridge between art and science” (McCorduck)
 “Tesler’s Theorem: AI is whatever hasn’t been
done yet.” (Hofstadter)
 “AI is a field of science and engineering
concerned with the computational understanding
of what is commonly called intelligent behavior,
and with the creation of artifacts that exhibit such
behavior.” (Shapiro)
 AI may be defined as the branch of computer
science that is concerned with automation of
intelligent behavior. (Luger & Stubblefield)
4

AI Application Areas
 Two fundamental AI research areas
 Knowledge Representation: represent the computer’s knowledge of
the world by some kind of data structures in the machine’s memory
 Search: a problem-solving technique that systematically explores a
space of problem states
 Game Playing
 Automated Reasoning and Theorem Proving
 Expert Systems
 Natural Language Understanding and Semantic Modeling
 Modeling Human Performance
 Planning and Robotics
 Machine Learning
 Neural Nets and Genetic Algorithms
13

Game Playing
 Games are good vehicles for AI research because
 most games are played using a well-defined set of rules
 board configurations are easily represented on a computer
 Games can generate extremely large search
spaces.
 Search spaces are large and complex enough to require
powerful techniques(heuristics) for determining what
alternatives to explore in the problem space.
14

Automated Reasoning and
Theorem Proving
 Automatic Theorem Proving is the oldest branch of AI.
 Theorem proving research was responsible for much of the early work
in formalizing search algorithms and developing formal representation
languages such as predicate calculus and logic programming language
PROLOG.
 Variety of problems can be attacked by representing the
problem description and relevant background information as
logical axioms and treating problem instances as theorems to
be proved.
 Reasoning based in formal mathematical logic is also
important.
 Many problems such as the design and verification of logic circuits,
verification of the correctness of computer programs, and control of
complex systems require automated reasoning.
15

Expert Systems(1)
 Expert systems are constructed by obtaining the
knowledge of a human expert and coding it into a
form that a computer may apply to similar
problems.
 domain expert provides the necessary knowledge of the
problem domain.
 knowledge engineer is responsible for implementing this
knowledge in a program that is both effective and intelligent
in its behavior.
16

Expert Systems(2)
 Many successful expert systems
 DENDRAL
 designed to infer the structure of organic molecules from their
chemical formulas and mass spectrographic information about
the chemical bonds present in the molecules.
 use the heuristic knowledge of expert chemists to search into
the very large possible number of molecular structures.
 MYCIN
 used expert medical knowledge to diagnose and prescribe
treatment for spinal meningitis and bacterial infections of the
blood.
 Provided clear and logical explanations of its reasoning, used a
control structure appropriate to the specific problem domain,
and identified criteria to reliably evaluate its performance.
17

Expert Systems(3)
 Many successful expert systems (Continued)
 PROSPECTOR
 for determining the probable location and type of ore deposits
based on geological information.
 INTERNIST
 for performing diagnosis in the area of internal medicine.
 XCON
 for configuring VAX computers.
18

Deficiencies of Current Expert
Systems
1. Difficulty in capturing “deep” knowledge of the
problem domain
 MYCIN lack any real knowledge of human physiology.
2. Lack of robustness and flexibility
3. Inability to provide deep explanations
4. Difficulties in verification
 may be serious when expert systems are applied to air
traffic control, nuclear reactor operations, and weapon
systems.
5. Little learning from experience
19

Natural Language Understanding
and Semantic Modeling(1)
 One of the long-standing goals of AI is the creation
of programs that are capable of understanding
human language
 Ability of understanding natural language seem to be one of the
most fundamental aspects of human intelligence
 Successful automation would have an incredible impact on the
usability and effectiveness of computers
 Real understanding of natural language depends on
extensive background knowledge about the domain
of discourse as well as an ability to apply general
contextual knowledge to resolve ambiguities.
20

Natural Language Understanding
and Semantic Modeling(2)
 Current work in natural language understanding
is devoted to finding representational formalisms
that are general enough to be used in a wide
range of applications.
 Stochastic models and approaches, describing
how sets of words “co-occur” in language
environments, are used to characterize the
semantic content of sentences.
21

Modeling Human Performance
 Design of systems that explicitly model some
aspect of human problem solving
 If performance is the only criterion by which a system will be
judged, there may be little reason to attempt to simulate
human problem-solving methods.
 Programs that take non human approaches to solving problems
are often more successful than their human counter parts
 Human performance modeling has proved to be a powerful
tool for formulating and testing theories of human cognition.
22

Planning and Robotics
 Planning attempts to order the atomic actions
which robot can perform in order to accomplish
some higher-level task.
 Planning is a difficult problem because of vast
number of potential move sequences and
obstacles.
 A blind robot performs a sequence of actions
without responding to changes in its environment
or being able to detect and correct errors in its
own plan.
23

Foundations of Artificial Intelligence
 AI foundations come from multiple disciplines:
 - Philosophy
 - Mathematics
 - Computer Science
 - Cognitive Science
 - Engineering

Philosophical Foundations
 - Logic and Reasoning (Aristotle, Turing)
 - Philosophy of Mind (Consciousness, Thinking)
 - Ethics and Responsible AI

Mathematical Foundations
 - Set Theory, Logic
 - Probability & Statistics
 - Linear Algebra & Calculus
 - Graph Theory

Computational Foundations
 - Algorithms & Complexity Theory
 - Search & Optimization
 - Automata & Formal Languages
 - Data Structures

Cognitive & Biological
Foundations
 - Neuroscience (Neural Networks) - study of brain and
nervous system
 - Cognitive Psychology (study of thought, learning and
mental organization)
 - Linguistics (NLP)
 - Behavioral Science (study of human and animal
behavior)

Engineering Foundations
 - Control Theory & Robotics - a field of applied
mathematics and engineering focused on the analysis and design of
systems that achieve a desired behavior through feedback mechanisms
 - Information Theory (Shannon) - a branch of applied
mathematics focused on quantifying, storing, and communicating
information
 - Cybernetics (Wiener) - the study of control and
communication in complex systems, encompassing both living
organisms and machines

State of the Art of Artificial
Intelligence (AI)
30

Machine Learning & Deep
Learning
 - Foundation Models (GPT-5, LLaMA 3, Gemini)
 - Self-Supervised Learning
 - Multimodal Models (text, image, video, audio)

Natural Language Processing
(NLP)
 - Large Language Models (LLMs) - AI models,
specifically deep learning models, trained on massive amounts of text
data to understand and generate human language
 - Contextual Reasoning & Long Memory - crucial
cognitive abilities that allow humans and machines to understand and
remember information within its surrounding context
 - Conversational AI & Assistants - Conversational AI
systems, often embodied as chatbots or virtual assistants, are designed
to facilitate human-computer interaction through text or voice, providing
personalized and dynamic responses based on user input and context.

Computer Vision
 - Vision Transformers (ViTs) - a type of deep learning
model that applies the transformer architecture, originally designed for
natural language processing, to computer vision tasks
 - Diffusion Models (Stable Diffusion, DALL·E 3) - a
type of generative model that create new data samples by learning to
reverse a process of gradually adding noise to existing data
 - Medical Imaging AI - Artificial Intelligence (AI) is
revolutionizing medical imaging by enhancing diagnostic accuracy,
efficiency, and personalized treatment planning

Reinforcement Learning (RL)
 - RL with Human Feedback (RLHF) - a technique used
to train AI models, particularly large language models (LLMs), by
incorporating human preferences into the training process
 - Robotics (manipulation, locomotion) - involves the
study and development of robots that can both move themselves
(locomotion) and interact with and manipulate objects in their
environment (manipulation)
 - Game AI (AlphaZero, MuZero) - AI system developed by
DeepMind that can learn to play complex games like chess, shogi, and
Go from scratch through self-play, without any human guidance or pre-
programmed knowledge of the games

Generative AI
 - Text-to-Image & Video (Sora, Runway) - Sora is
OpenAI's video generation model, designed to take text, image, and
video inputs and generate a new video as an output. Runway is
a global AI research and technology company building foundational AI
research models and tools to create multimodal simulators of the world.
 - Code Generation (GitHub Copilot X, Code
Llama) - Code generation with AI models like GitHub Copilot X and
Code Llama represents a significant advancement in software
development, aiming to enhance developer productivity and efficiency
 - Synthetic Data for Training - Synthetic data is generated
by AI trained on real world data samples.

AI for Science & Healthcare
 - Protein Folding (AlphaFold2, RoseTTAFold) -
AlphaFold2 models and RoseTTAFold models are a good indicator of
protein foldability.
 - Clinical Decision Support – technology that helps
clinicians takes data-driven decisions
 - Climate & Physics Simulations - utilize complex
mathematical models and powerful computers to represent and predict
the behavior of Earth's climate system

Ethics, Safety & Regulation
 - AI Alignment & Bias Reduction - AI alignment is the
process of encoding human values and goals into AI models to make
them as helpful, safe and reliable as possible
 - AI Governance (EU AI Act 2024) - world's first
comprehensive legal framework for regulating artificial intelligence
 - Explainable AI (XAI) - refers to methods and techniques that
allow humans to understand, trust, and manage AI systems

Summary
 AI in 2025 is driven by:
 - Foundation & Multimodal Models
 - Generative AI & Healthcare Applications
 - Ethical Governance & Future Innovations

Intro AI DBATU.ppt ppt useful engineering

More Related Content

Similar to Intro AI DBATU.ppt ppt useful engineering

Recently uploaded

Intro AI DBATU.ppt ppt useful engineering