A warning (I) � CPSC 433 - Artificial Intelligence � If you do not like or have problems with the following Fall 2017 � topics and examples, then you should consider dropping this course: � n Logical formulas, logical thinking � If you hated your logic course, then you will hate this course (and probably fail it) � Jörg Denzinger � ICT 752 � n Mathematical modeling � n Very abstract thinking � denzinge@cpsc.ucalgary.ca � (away from applications, concrete programs) � http://www.cpsc.ucalgary.ca/~denzinge/ � CPSC 433 - Artificial Intelligence Jörg Denzinger A warning (III): � A warning (II): Logic � Mathematical modeling � Traveling salesman problem with n cities. � n Formulas like � � Prob = {(1,c 2 ,...,c i ) | c i ∈ {2,...,n}, 1 ≤ i ≤ n}. � ∀ x. ∃ y. (P(x) → ¬ Q(y)), p ∧ (q ∨ ¬ r), � � Erw(((1,...,c n-1 ),?), ((1,...,c n-1 ),yes)). � ∀ x. ∀ y. EQ(f(x),g(x,y)) � � Erw(((1,...,c i ),?), ((1,...,c i ),yes)), � n Calculus rules like � � if f bound ((1,...,c i )) is larger or equal to the best � C ∨ P , D ∨ ¬ P ’ � If σ (P) ≡ σ (P ’ ) � � solution to the problem instance found so far. � σ (C ∨ D) � � Erw(((1,...,c i ),?), � � ((1,...,c i ),?,((1,...,c i ,c ’ i+1,1 ),?),...,((1,...,c i ,c ’ i+1,k ),?))), � � if k = n-i, {c ’ i+1,1 ,...,c ’ i+1,k } = {2,...,n} - {c 2 ,...,c i }, � � c ’ i+1,j < c ’ i+1,p if j < p. � and so on � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger � Grading � Other organizational items � n Exam component: � n Office hours: Tuesdays, Thursdays 14:00 - 15:00 � � a: � Midterm � � 20% � individual � or by appointment � � b: � Final � � � 30% � individual � (make use of the "by appointment"!) � n Assignment component � n Course web page: � � c: � Paper presenting 2 solutions � 20% � team � http://www.cpsc.ucalgary.ca/~denzinge/courses/ � � d: � Implemented System/Demo � 30% � team � � � � � cs433-fall2017.html � n Both components must be passed � n Midterm: � � (2a + 3b)/5 has to be greater than 0.5 � � October 31, 2017, in class (50 min), room tbd � n You will get grades for each item above: � n Labs: Start next week. � Course grade = (2a + 3b + 2c + 3d)/10 � Participation in first lab absolutely required! � � rounded to nearest grade level � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger 1
Artificial Intelligence - � Example - � An example of what we face � What is involved (I) � “ Can you get me a coffee, please ? (or !?) ” � Tasks: � n Hear the sentence and transfer it in a written version � n Parse and “ understand ” the written version � (i.e. transfer it into an internal representation) � F Natural language understanding � F Knowledge Representation � n Use sensory input to identify the current situation and possibilities � F Vision � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger Example - � Example - � What is involved (II) � What is involved (III) � n Define a goal; learned information about user (he n Move to waiter � likes his coffee with cream and sugar) is necessary � F Robotics (but also Vision, (Re-)Planning) � F Machine Learning � n Deal with waiter � F Deduction � F Natural language generation � n Planning what to do: � F Human-computer interaction � ● Move to coffee bar � F Multi-Agent Systems (if waiter is or has a robot) � ● Interact with waiter � ● Prepare coffee (add cream and sugar; stir) � ● Bring coffee to user � F Planning � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger Example - � and we do not want � So, what is Artificial Intelligence (AI) � AI deals with the development of systems either displaying a behavior humans associate with intelligence or solving a problem humans think only an intelligent being can solve. � F subjective definition � F changes over time! � I prefer to use the term Knowledge-based Systems � F describes what systems have to deal with and do: � knowledge and knowledge processing � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger 2
AI - � AI - � The Beginning � The Past (I) � Goals : Develop computer programs that show an Goals : Select a field in AI and an application area and intelligent behavior and that can serve as companion “ make it work ” � to human beings. � � Example : the GPS program (General Problem Solver) Examples : � “ describe problem in logic and it will solve it using n Natural language understanding and dictating � logical inferences ” � n Expert systems for configuration and diagnosis in Result : Complete failure and enormous negative small application areas � reaction in media (and elsewhere) � n Playing a game � Problems : One knowledge representation scheme with one knowledge processing mechanism not enough + n Scoring high in intelligence tests � search spaces are enormous. � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger AI - � AI - � The Past (II) � The Present (I) � New developments and Applications: � Result : It worked within the expectations � n Multi-Agent Systems: � � ● Cooperation concepts to bring together the many Problems : � specialized systems to tackle harder tasks � n Very different methods used in different fields � ● User has many special (and otherwise rather n Application areas very specialized � dumb) assistants � n Systems resemble “ idiot savante ” (person without n The Internet � any interest outside his/her field) � ● Requires agents for “ intelligent ” routine tasks that n Hard problems still a problem � are specialized and limited � ● “ Real world ” is inside the computer � F environment is AI friendly � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger AI - � AI - � The Present (II) � The Present (III) � n Service-based computing: � n Self-X Systems: � ● Offer functionality at your web site and allow ● Started as IBM ’ s (and now others) code for others to use it � learning/adaptive systems � ( F describe service and how to access it, best in a ● Sees more and more areas within CS that want to form understandable by a program) � include learning/adaptation into their concepts ● Decentralization of knowledge representation and and systems � processing � ● Self-organization, self-healing, … � ● Tries to rescue the semantic web idea � ● General goal: make systems easier to configure, maintain and adapt to usage � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger 3
AI - � AI - � The Present (IV) � The Future ? (I) � n Areas in AI grow together: � n Specialized agents (avatars) helping and representing ● Personal assistants like Alexa: restricted natural users (in the new digital world, i.e. moving Siri/ language understanding, machine learning, Alexa/... forward) � (intelligent) search techniques in the Internet and n “ Command-and-Control ” systems for management use of service-based computing � and CEOs in large companies � ● Watson: dedicated knowledge representation for F Management by exception � an application area, machine learning to add to n Cooperating intelligent robots in disaster areas and represented knowledge (often using natural hostile environments represent humans acting out of language understanding) and restricted inference save environments � mechanisms � n Automation of trade and commerce � ● ... � F Management by Policies � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger AI - � AI - � The Future ? (II) � The (little) Farther Future ? � In general: � Together with advances in many engineering disciplines, have AIs (embodied or not) replace Use of AI enhances what a human being can do humans in specific areas (that require only dedicated intellectually (assistants) � and thus limited knowledge): � Still the hard problems: � n driving cars/trucks/trains/planes/... � n Achieving necessary flexibility to interact with n “business drones” � human beings in real world � n factory workers � n Find concepts to combine the methods of the n ... � different fields � n Common sense reasoning � CPSC 433 - Artificial Intelligence Jörg Denzinger CPSC 433 - Artificial Intelligence Jörg Denzinger Goals of this course � n Introduction of basic concepts and techniques � n Understanding the need to be precise and learn how to be precise � n Understanding basic problems and limitations � n Knowing advantages and disadvantages of a particular concept or technique (in relation to others) � n Allows you a critical view on what people consider AI to be able to do � CPSC 433 - Artificial Intelligence Jörg Denzinger 4
Recommend
More recommend
