CS 730/830: Intro AI
What is AI?
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 1 / 23
Prof. Wheeler Ruml
TA Tianyi Gu
“Thinking inside the box.”
5 handouts: course info, project info, schedule, slides, asst 1
sign up sheet/laptop (grading email, piazza)
What is AI?
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 2 / 23
My Definition of AI
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 3 / 23
What is a Robot?
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 4 / 23
Artificial physical system that takes adaptive action.
■ remote-controlled car■ power tool■ robotic surgery■ motion sensor■ thermostat■ anti-lock brakes■ automated delivery■ autopilot■ self-driving car■ Ava, Data. . .
What is Intelligence?
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 5 / 23
What behaviors require intelligence?What makes an agent intelligent?
Different Goals in AI
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 6 / 23
How to understand Intelligence?
Cognitive modeling: behaves like a humanEngineering: achieve human performanceRational: behaves perfectly, normativeBounded-rational: behaves as well as possible
Subfields: knowledge representation and reasoning, computerproblem-solving, planning, machine learning, natural languageprocessing, (autonomous) robotics, intelligent agents,multi-agent systems, distributed AI, intelligent user interfaces,machine visionOther terms: computational intelligenceRelated: adaptive behavior, complex adaptive systems, artificiallife, cognitive modeling
Relations
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 7 / 23
■ CS: algorithms■ Engineering: applications■ Cognitive psychology: modeling■ Philosophy: mind, rationality■ Math: logic, statistics■ Linguistics: language processing■ Operations research: optimization■ Economics: agents, incentives
AI Today
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 8 / 23
■ Game playing: chess, checkers, backgammon, Jeopardy!,crosswords, go
■ Design: VLSI, jet engines■ Diagnosis: POS, NASD, loans, customer service, medical
testing and classification, DS1■ Planning: airports, flight routes, Dell, DART■ Learning: Amazon, Netflix, Walmart, Facebook■ Robotics: ping-pong, beer fetch, driving, flying■ Language: voice recognition, translation■ Vision: scene descriptions, face recognition■ Hidden: logistics, data center control, distribution centers
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
Honda Asimo: virtually no autonomy.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
NASA Mars Science Lab: some navigation autonomy.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
NASA Deep Space 1: temporarily self-commanded.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
AUVs: dynamic environment, poor communication.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
Boston Dynamics LS3: follow me.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
Kiva Systems: bring inventory to pickers.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
KAIST Hubo: winner of the 2015 DRC.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
Willow Garage PR2: 22 degrees of freedom.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
Yamaha RMax at Linkoping University: autonomous.
Robots Today: Beautiful Hardware
What is AI?
■ My Definition
■ Robots
■ Intelligence
■ The Goal
■ Relations
■ AI Today
■ Robots Today
This class
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 9 / 23
Google Self-Driving Car: over 1.8M miles, 13 minor accidents.
This class
What is AI?
This class
■ The AI View
■ An AI Agent
■ Schedule
■ Course Mechanics
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 10 / 23
The AI View of An Agent
What is AI?
This class
■ The AI View
■ An AI Agent
■ Schedule
■ Course Mechanics
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 11 / 23
The AI View of An Agent
What is AI?
This class
■ The AI View
■ An AI Agent
■ Schedule
■ Course Mechanics
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 11 / 23
percepts → → actions
An AI Agent
What is AI?
This class
■ The AI View
■ An AI Agent
■ Schedule
■ Course Mechanics
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 12 / 23
agent
world
actions
sensing
An AI Agent
What is AI?
This class
■ The AI View
■ An AI Agent
■ Schedule
■ Course Mechanics
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 12 / 23
world model
planner
agent
world
actions
sensing
An AI Agent
What is AI?
This class
■ The AI View
■ An AI Agent
■ Schedule
■ Course Mechanics
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 12 / 23
world model
planner
search
agent
world
actions
sensing
Schedule
What is AI?
This class
■ The AI View
■ An AI Agent
■ Schedule
■ Course Mechanics
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 13 / 23
1. planning: vacuum tasks, hovercraft motion, puzzlestate-space searchconstraint satisfactioncombinatorial optimization
2. KR: theorem proverspropositional logicfirst-order logic
3. more planning: general planner, probabilistic plannerdomain-independent planningMarkov decision processes
4. perception: digits, shapes, localizationsupervised and unsupervised learninghidden Markov models
See also: Intro to mobile Robotics, Intro to Machine LearningNot: NLP, cognitive modeling, philosophy
Course Mechanics
What is AI?
This class
■ The AI View
■ An AI Agent
■ Schedule
■ Course Mechanics
Problems in AI
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 14 / 23
■ General information■ Schedule■ Project■ Asst 1■ Names
Problems in AI
What is AI?
This class
Problems in AI
■ Agent Designs
■ Examples
■ Environments
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 15 / 23
Agent Designs
What is AI?
This class
Problems in AI
■ Agent Designs
■ Examples
■ Environments
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 16 / 23
Agent ⇔ Environment
Perception: vision, state estimationPlanning: low/high-level, on/off-line, incremental/repairActing: dispatching, monitoring, diagnosis
Reflex: sensors → effectorsReflex with state: sensors + state → effectors + new stateGoal-based: reason from goals to meansUtility-based: use quantitative measure of happiness
What kind of agent?
What is AI?
This class
Problems in AI
■ Agent Designs
■ Examples
■ Environments
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 17 / 23
1. Thermostat2. autonomous armed drone3. Mail delivery robot4. Medical diagnosis system
Environments
What is AI?
This class
Problems in AI
■ Agent Designs
■ Examples
■ Environments
Search
Wheeler Ruml (UNH) Lecture 1, CS 730 – 18 / 23
Observability: complete, partial, hiddenPredictability: deterministic, strategic, stochasticInteraction: one-shot, sequentialTime: static, dynamicState: discrete, continuous (also time, percepts, and actions)Agents: single, multiagent (competitive, cooperative)
State-Space Search
What is AI?
This class
Problems in AI
Search
■ Contents
■ Cognitive Science
■ A Search Space
■ EOCQs
Wheeler Ruml (UNH) Lecture 1, CS 730 – 19 / 23
Contents
What is AI?
This class
Problems in AI
Search
■ Contents
■ Cognitive Science
■ A Search Space
■ EOCQs
Wheeler Ruml (UNH) Lecture 1, CS 730 – 20 / 23
This particular pattern of molecules known as a ’human being’has evolved an amazing depth of consciousness: an ability tointernally model the reality beyond the senses, to imagine futuresthat have never happened, to use language, to use rationality tobuild and test theories about our universe, to become self-aware.—Jeff Lieberman (artist, roboticist)
Cognitive Science
What is AI?
This class
Problems in AI
Search
■ Contents
■ Cognitive Science
■ A Search Space
■ EOCQs
Wheeler Ruml (UNH) Lecture 1, CS 730 – 21 / 23
The ability to think is perhaps the most distinctive of humancapacities. Typically, thinking involves mentally representingsome aspects of the world (including aspects of ourselves) andmanipulating these representations or beliefs so as to yield newbeliefs, where the latter may aid in accomplishing a goal.—Edward E. Smith (Psychology, U Michigan)
The ability to solve problems is one of the most importantmanifestations of human thinking. ... We might thereforesuspect that problem solving depends on general cognitiveabilities that can potentially be applied to an essentiallyunlimited range of domains.—Keith Holyoak (Psychology, UCLA)
A Search Space
What is AI?
This class
Problems in AI
Search
■ Contents
■ Cognitive Science
■ A Search Space
■ EOCQs
Wheeler Ruml (UNH) Lecture 1, CS 730 – 22 / 23
EOCQs
What is AI?
This class
Problems in AI
Search
■ Contents
■ Cognitive Science
■ A Search Space
■ EOCQs
Wheeler Ruml (UNH) Lecture 1, CS 730 – 23 / 23
Please write down the most pressing question you have aboutanything related to the course (no need to include your name)and put it in the box on your way out.Thanks!