18
CS 4620 Intelligent Systems

CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Embed Size (px)

Citation preview

Page 1: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

CS 4620Intelligent Systems

Page 2: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

What we want to do today

• Course introductions• Make sure you know the schedule for the next

three weeks

Page 3: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

How 4620 is different from 3610

• 3610– Was about a teacher and a textbook teaching you

about a number of specific topics– Was mainly about general techniques of AI which we

applied in small, “toy” problems.• 4620– Will focus on you teaching yourself and your

classmates on topics that will depend on you.– Will focus on a “large” systems of knowledge and

managing that knowledge to actually do something

Page 4: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Course Goals

• The first goal, is for you to explore the idea of intelligent systems in some depth. – In this course, we will define intelligence as "the

capacity to acquire and apply knowledge" thus intelligent systems are those which modify their actions based on prior interactions.

– In this course you will study the techniques and concepts common to this sub-discipline of AI by constructing a non-trivial intelligent system.

Page 5: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Course Goals

• The second goal of the course is to provide a capstone experience to your undergraduate computer science curriculum. – To this end, the course provides an opportunity for

you, as a part of a software development team in a studio format in conjunction with your instructor, to construct a large software system.

– The concept of a software studio may be new to you. You should make yourself familiar with the online description of this process.

Page 6: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Software Studio Quick Intro

• Takes it’s main idea from the concept of an art or architecture class taught in studio.– Students are given an assignment.– The teacher moves around asking and answering

questions.– After a certain length of time some or all of the students

gather for a “design critique”• Gather around one student’s project. • That student explains what they did, and answers questions from

his/her peers about why they did it that way.• Meant to be a chance for students to REFLECT on their PRODUCT

as well as the PROCESS.

Page 7: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Software Studio Quick Intro

• We will blend that idea with “Agile Software Development with Scrum”

• Once we get started with the project you will complete five, two-week cycles of development and testing (called Sprints)

• Most sprints will be followed by a design crit.

Page 8: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Daily Formats

• Guided Practice (GP) – These are somewhat related to the "free work days" you

might have in other courses. However, there are some differences and that is why I don't use that term.

– On GP days you and your team are expected to be in class and working on your project.

– I will frequently move around asking questions and expect that you should feel free to ask questions of both me and your classmates.

– You will receive points based on how well you are engaged with your teammates, classmates, and the projects.

Page 9: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Daily Formats

• Design Crits (DC) – If we were going to compare this to some of the other courses

you have had then these might be labeled "presentation days." – But it isn’t as PASSIVE as “presentation days” in other courses.– On DC days one or more teams will be expected to explain to

the rest of the class where their project is at and/or where their project is going.

– Members of the audience should be actively engaged in asking the other groups questions about decisions and helping consider problems.

– You will receive points based on your participation from BOTH sides of the podium

Page 10: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Daily Formats

• Readings Discussion (RD)– In the early part of the semester we will spend several

class sessions discussing readings which will be selected to give you an introduction to the idea of intelligent/expert systems based on the course needs.

– Most of these discussion will be "led" by the student members of the class, not the instructor.

– You will receive points based on your preparation to participate in a meaningful discussion on the readings.

Page 11: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Grading

Page 12: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

For Wednesday

• Read materials about the course on the class website ( www.cs.uni.edu/~schafer/4620 )– Syllabus– Information about the Software Studio Approach– The deliverable schedule for the course project

• Read two “chapter one” chapters from two different textbooks we COULD have used for this type of course:– PDF files posted online

Page 13: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

For Friday

• Come with ideas for projects

Page 14: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

For week #2

• Form teams• Consider techniques from AI that might be

beneficial• Begin to do general knowledge acquisition for

your project

Page 15: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

For week #3

• General knowledge acquisition for your project.

• Creating an initial project analysis– Focus on WHAT rather than HOW at this point

Page 16: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

For week #4 and 5

• We will read about topics that could be of use to you in your projects (which topics will depend on you).

• You will work outside of class as a team to develop a FORMAL plan for the project

Page 17: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Types of projects you might consider

• Diagnostic system (take in information about a situation and diagnose what is wrong/what to test next)

• Filtering System (Classify images, text, email, etc. Preferably one which learns)

• Design systems (know rules for construction of an object and propose new objects to generate given the rules)

• Design aids (Merge of several of above. As the human does the design, suggests next steps)

• Schedulers/Planners for any number of domains (faculty course scheduler, lab worker scheduler, etc)

• Learning Agent for a particular game• Intelligent tutoring system.

Page 18: CS 4620 Intelligent Systems. What we want to do today Course introductions Make sure you know the schedule for the next three weeks

Examples of previous projects

• Agents to play a variety of board and computer games– Poker– Super Mario Brothers

• Medical diagnostic system with a learning component

• Neural network based OCR• NLP story problem solver