CS561Lecture1

CS561Lecture1 - Foundations of Artificial Intelligence...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Foundations of Artificial Intelligence Lecture 1: Overview and Background (Part I, Chapter 1) Spring 2010 Instructor: Paul S. Rosenbloom TA: Hyokyeong Lee Today’s Lecture  Intelligence, Agents & Artificial Intelligence  Examples and Grand Challenges  Where AI fits in computing  Course Details 2 What is Intelligence? (1)  What is measured by a test/ standard, e.g.:  “Intelligence is what is measured by intelligence tests.” E. Boring  Thought processes, or behavior, that is indistinguishable from what a human would produce (at some level of abstraction)  The Turing test 3 What is Intelligence? (2)  A conglomeration of specific capabilities, or what is common among them, usually based on humans as prototypes, e.g.:  “The general mental ability involved in calculating, reasoning, perceiving relationships and analogies, learning quickly, storing and retrieving information, using language fluently, classifying, generalizing, and adjusting to new situations” (Columbia Encyclopedia)   “… a very general mental capability that, among other things, involves the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly and learn from experience.” (Editorial in Intelligence with 52 signatories) 4 What is Intelligence? (3)  A specific capability, e.g.:  “The capacity to acquire and apply knowledge.” (The American Heritage Dictionary)  “The ability to plan and structure one’s behavior with an end in view.” (J. P. Das)  “… the ability of an organism to solve new problems …” (W. V. Bingham)  “The capacity to learn or to profit by experience.” (W. F. Dearborn)  “The ability to carry on abstract thinking.” (L. M. Terman)  “… ability to achieve goals in a wide range of environments.” (S. Legg & M. Hutter)  … ability to act rationally; that is, “does the ‘right thing,’ given what it knows.” (S. Russell & P. Norvig) 5 Systems of Interest  Have goals to achieve  May concern internal or external situations  May be endogenous or exogenous  Have capabilities to perceive and act in service of their goals  For external environments, might include eyes, ears, hands, legs, etc.  Or wheels, laser range finders, etc.  Can embody “knowledge” concerning their goals, capabilities, and situations 6 Agents  These systems are generally called Agents (or Intelligent Agents) within AI  Differs from notion of agent in Hollywood and in the rest of CS, where the focus is on proxies (or representatives)  Differs also from how the book minimally defines an agent  The textbook is constructed around rational agents as the organizing theme for AI  One way to bring it all together  However, most of the field, and much of the book focuses on particular aspects/parts 7 Some Relevant Agent Aspects  Generality: Scope of goals that can be attempted and capabilities that can be brought to bear on them  Can the agent play both chess and tennis?  Can it solve math problems and drive a car?  Does it have the capability to successfully perform all of the tasks of which an adult human is capable?  Requires scope in both mental (Parts II-V) and perception/action (Part VI) capabilities  Literacy: Extent of knowledge available  Ignorance by itself is not lack of intelligence  Requires ability to acquire (Part V) and represent knowledge (Part III and bits of others) 8 Relevant Aspects (Cont.)  Rationality: Making best decisions about what to do given goals, knowledge and capabilities  Thermostats may exhibit perfect rationality, but exhibit highly restricted generality  Requires problem solving (Part II), reasoning and planning (Parts III-IV)  Autonomy: Operating without assistance  Requires ability to perceive and act for self (Part VI), ability to learn to perform new tasks well (Part V), plus other things, such as motivations, emotions, selfmonitoring and error recovery  Collaboration: Working well with others  Requires communication (Part VI) plus other things, such as understanding others and acting as part of group 9 What is Intelligence? (4)  The common underlying capabilities that enable a system to be general, literate, rational, autonomous and collaborative  Generally forms what is known as the Cognitive Architecture  There may be more involved as well, but this is a start  Although the textbook focuses on rationality in its definition of intelligence, it does not neglect generality or literacy in its coverage of AI   Autonomy and collaboration are even less of a focus, but still do show up in various ways 10 The Study of Intelligence  Cognitive Science is the interdisciplinary study of mind and intelligence in both natural and artificial systems  Textbook mistakenly limits it to natural systems  Disciplines involved include  Philosophy: Questions, concepts and formalisms  Psychology: Data and theories about natural systems  Linguistics: Study of language structure and use  Neuroscience: Data/theory that ground mind in brain  Anthropology: Intelligence in/across context/culture  Sociology: Data/theory on natural societies  Computer science: Study and construction of artificial systems, plus methods for modeling natural systems1 1 What is Artificial Intelligence (AI)?  Some Bad definitions  “The study of how to make computers do things at which, at the moment, people are better.” (E. Rich & K. Knight)  “The concept of making computers do tasks once considered to require thinking.” (Medford Police)  The use of logic or rules or … to solve problems on computers.  “An algorithm by which the computer gives the illusion of thinking like a human.” (D. Gruber)  “Making computers behave like humans.” (Webopedia) 12 A Better Definition  “The scientific understanding of the mechanisms underlying thought and intelligent behavior and their embodiment in machines.” (AAAI)  Overlaps strongly with Cognitive Science and its various subdisciplines, but also relates to:  Mathematics: Formalizations and analysis  Economics: Decision making  Operations research: Optimization and search  Engineering: Robotics 13 Some Examples DARPA Urban Challenge  “Autonomous vehicles that safely execute missions in a complex urban environment with moving traffic.”  “The objective of this program is safe and correct autonomous driving capability in traffic at 20 mph.” 15 Capabilities to be Demonstrated (1)   Complete a mission defined by an ordered series of checkpoints in a complex route network.   Interpret static lane markings (e.g., white and yellow lines) provided with the route network definition file and behave in accordance with applicable traffic laws and conventions.   Exhibit context-dependent speed control to ensure safe operation, including adherence to speed limits.   Exhibit safe-following behavior when approaching other vehicles from behind in a traffic lane. This includes maintaining a safe-following distance.   Exhibit safe check-and-go behavior when pulling around a stopped vehicle, pulling out of parking spot, moving through intersections, and in situations where collision is possible.   Stay on the road and in a legal and appropriate travel lane while en route, including around sharp turns, through intersections, and while passing.   Navigate safely in areas where GPS signals are partially or entirely blocked.   Follow paved and unpaved roads and stay in lane with very sparse or low accuracy GPS waypoints. 16 Capabilities to be Demonstrated (2)   Change lanes safely when legal and appropriate, such as when passing a vehicle or entering an opposing traffic lane to pass a stopped vehicle.   Merge safely with traffic moving in one or more lanes after stopping at an intersection.   Pull across one lane of moving traffic to merge with moving traffic in the opposing lane.   Stop safely within 1 meter of the stop line at a stop sign intersection and proceed without excessive delay (less than 10 seconds) according to intersection precedence rules.   Exhibit proper queue behavior at an intersection, including stopping at a safe distance from other vehicles and stop-and-go procession to the stop line without excessive delay.   Navigate toward a destination in a large, open area where minimal or no GPS points are provided, as in loading dock areas or parking lots.   Safely pull into and back out of a specified parking space in a parking lot.   Safely execute one or more three-point turning maneuvers to effect a U-turn.   Dynamically re-plan and execute the route to a destination if the primary route is blocked or impassable. 17 Capabilities Beyond the Scope of the Program   Recognition of external traffic signals such as traffic lights and stop signs, through the use of sensors.   Recognition of pedestrians and pedestrian avoidance.   Behaviors necessary for highway driving such as high speed passing or high speed merge at an onramp.   Speed limits for the Urban Challenge will be 30 mph or less.   Driving in difficult off-road terrain is outside the scope of the program.   Off-road navigation in an unpaved area, travel along roads with potholes, and travel along a dirt road are within scope. 18 DARPA Urban Challenge Summary Video 19 Deep Blue  In 1997 Deep Blue became the first machine to win a match against a reigning world chess champion (by 3.5-2.5) 20 Deep Blue Combined  Parallel and special purpose hardware"  A 30-node IBM RS/6000, enhanced with "  480 special purpose VLSI chess chips"  A heuristic game-tree search algorithm"  Capable of searching 200M positions/sec"  Searched 6-12 moves deep on average, sometimes to 40"  Chess knowledge"  An opening book of 4K positions and 700K GM games"  An endgame database for when only 5-6 pieces left"  A positional evaluation function with 8K parts and many parameters that were tuned by learning over thousands of Master games 21 Deep Blue Trailer 22 Tactical Language   Teach language and culture in context   Trainee plays one character   Teammates & locals are “socially intelligent” agents   Understand task and other characters   Have goals and emotions   Use speech and gesture   One teammate acts as guide   Versions exist for Iraqi, Pashto, French, Dari, … 23 Tactical Iraqi Movie 24 Grand Challenges for AI  Human-level AI  Intelligent virtual humans and humanoid robots  Superhuman performance in limited domains  Beat the world champion at chess (or Go)  Effective control of very complex systems  Specific advanced capabilities of interest  Mathematical or scientific discovery  Automatically drive a vehicle in real world  Read and answer questions in a textbook  Autonomously behave and learn continuously over years  Automated real-time language and speech translation  Learn to perform a new task from scratch 25 Relational Architecture for Computing Theory, Algorithms Physical Mechanical, Optical, Electronic, Quantum & Chemical Computing Modeling/Simulation, Data Bases/Systems, Digital Physics Sensors, Scanners, Computer Vision, Optical Character Recognition, Localization Locomotion, Fabrication, Manipulation, OpenLoop Control Robots, Closed-Loop Control Social Wizard of Oz, Mechanical Turks, Human Cognition Life Genomic, Neural, Immunological & Evolutionary Computing Artificial Life, Biomimetics, Systems Biology Eye, Gesture, Expression & Movement Tracking, Biosensors Bioeffectors, Haptics, Sensory Immersion Brain Computer Interfaces Computing Compilers, OS, Emulation, Reflection, Abstractions, Procedures, Architectures, Languages Networking, Security, Parallel Computing, Distributed Systems, Grids Implementation Implemented Implements Artificial Intelligence, Cognitive Modeling, Virtual Humans, Autonomic Systems Mice/Keyboards, Learning, Programming, User Modeling, Authorization, Speech Understanding Screens, Printers, Graphics, Speech Generation, Cognitive Augmentation Human Computer Interaction, Full Immersion, Games Influenced Interaction Influences Bidirectional Systems Architecture for Computing 27 Course Details  Course Schedule: TTh 12:30 – 1:50  Course Location: ZHS 352  Course Format: Primarily lecture with questions and discussion  Course Information: Blackboard  Syllabus can also be found at http://cs.usc.edu/~rosenblo/Spring-10-CS561Syllabus-PR.htm  Textbook: Artificial Intelligence: A Modern Approach (3rd Edition), by Russell, S. J. & Norvig, P., Prentice Hall, 2009  Best if you can read relevant material before class  Lectures will be posted on Blackboard after class 28 Teaching   Instructor: Paul Rosenbloom   Office: SAL 238   Email: rosenbloom@usc.edu   Phone: (213) 740-4780   Office Hours: TTH 2:00-3:00   TA: Hyokyeong Lee   Office: PHE 316   Email: hyokyeol@usc.edu   Phone: (213) 740-5351   Office Hours: W 10:00-12:00 29 Topics  Introduction [1/12-1/14]  Problem Solving [1/19-2/2]  Knowledge and Reasoning [2/7-3/6]  Reasoning, Planning and Knowledge [2/4-3/9]  Uncertainty [3/11-3/30]  Learning [4/8-4/20]  Miscellaneous and Wrap up [4/22-4/29] May evolve as I continue to adapt to new edition 30 Grading   Programming Projects {30%} Expected to be in C++ 1.  Problem solving [2/4-2/23] {10%} 2.  Planning/KRR [4/1-4/26] {20%} Have 2 late days for semester, after which lose significant fraction of credit for each day late   Midterms {40%} 1.  [2/16] {20%} 2.  [4/6] {20%}   Final [5/12, 2-4PM] {30%} Exams are open book & notes, but all work (including projects) must be your own! 31 ...
View Full Document

Ask a homework question - tutors are online