Richard Anderson
Department of Computer Science and Engineering
University of Washington
Dec 15, 2008
Microsoft Learning
1
•
How can computing technology enhance
education?
– Focus on classroom instruction
•
Challenges:
–
–
–
–
Extending reach of education
Increasing interaction
Addressing problems of scale
Facilitating expression of ideas
Dec 15, 2008
Microsoft Learning
2
Video conferenced
distance education
Presentation
systems
UW PMP
DISC
Classroom
Presenter 2.0
Classroom
interaction systems
Tutored Video
Instruction
Classroom Feedback
System
UW CC TVI
Project
CATs for CS1
ConferenceXP
Center for
Collaborative
Technologies
Beihang TVI project
Classroom
Presenter 3.0
Structured Interaction
Presentations
(SIP)
Digital StudyHall
Student submissions
with CP
Dec 15, 2008
Microsoft Learning
3
•
Deployment driven
– Classroom use
– Technology development and promotion
•
Goals and success criteria
– Adoption of technology and methodology
– Influence educational practice
•
Target specific deployments
– Innovate in some aspect of deployment
Dec 15, 2008
Microsoft Learning
4




Distance Learning and Video Conferenced
Classes
Tutored Video Instruction
Digital Study Hall
Educational Technology for Low Resource
Environments
Dec 15, 2008
Microsoft Learning
5
•
•
Multi-site internet based audio-video
conferencing
UW Master’s Program
– Site-to-site courses between UW and Microsoft since
Winter 1997
– www.cs.washington.edu/education/dl/course_index.html
– Master’s level courses
– Goal: interaction across sites
• Approximate single classroom
– Various technologies have been used since the
program was introduced
Dec 15, 2008
Microsoft Learning
6
Dec 15, 2008
Microsoft Learning
7
Dec 15, 2008
Microsoft Learning
8
•
Initial phase
• Winter 1997 – Winter 2002
– Polycom + Netmeeting for
•
PPT and SmartBoard
MSR DISC Project
– Target: UW, CMU, UCB,
•
Brown graduate class
– Spring 2002
ConferenceXP
– Since Spring 2003
– Four way courses, Autumn
2004, Autumn 2005,
Autumn 2006
• UW, MSR, UCB, UCSD
Dec 15, 2008
Microsoft Learning
9
•
•
High quality, low latency video to support
interactive classes
High bandwidth internet video conferencing
– Internet2
– Multicast
•
Single machine deployment
– High end PC
– Performance limit: handling multiple high resolutions
video streams
•
Innovative presentation tools
Dec 15, 2008
Microsoft Learning
10
•
What went wrong
– Technology and systems
–
–
–
–
–
–
–
•
failures
Multicast networking
High cost of interruptions
Audio
Loss of trust
Room configuration issues
Lack of control of lecture
room
Production quality
•
How to Fail at
VideoConferenced Teaching
– Microsoft Faculty Summit 2002
– Anderson & Beavers
Meta lesson
– Learn more from failures
than from successes
Dec 15, 2008
Microsoft Learning
11

Goals
 Real time interaction between sites
 High quality video

Challenges
 High bandwidth connections
▪ Multicast vs. Unicast
 Classroom Audio
 Establishing a pattern of interaction
Dec 15, 2008
Microsoft Learning
12

March 29, 2008,
LACCIR Meeting
 Latin American and
Carribbean
Collaboration for ICT
Research
Seattle and University
of Chile, Santiago,
Chile
 Seminar presentation

 October 15, 2008

CXP Unicast reflector
Dec 15, 2008
Microsoft Learning
13

Masters class
 University of
Washington
 Lahore University of
Management Science
 Microsoft

Computing for the
Developing world
Dec 15, 2008
Microsoft Learning
14

Ensuring adequate bandwidth
 Limited bandwidth to Pakistan
 Reliability
 Multicast
 Ensuring this did not compromise UW-MS class
 Limited time to prepare
Dec 15, 2008
Microsoft Learning
15
CXP
Video
cameras
Video
cameras
Audio
Audio
PMP
VENUE
Video
Displays
Video
Displays
Speakers
Speakers
UW
Microsoft
Archiver
Student Tablets
CP3
Display
Dec 15, 2008
Student Tablets
CP3
Instructor
Microsoft Learning
CP3
CP3
Display
16
CP3
CP3
Microsoft
PMP
Venue 1
LUMS
PMP
Venue 2
Archiver
CP3
CP3
Server
Dec 15, 2008
UW
Microsoft Learning
17
Tablet PC based
presentation and
classroom interaction
system
 Ink based presentation
 Classroom Activites

Dec 15, 2008
Microsoft Learning
18
Student
Instructor
Public
Display
Dec 15, 2008
Student
Microsoft Learning
19
Dec 15, 2008
Microsoft Learning
20

High connectivity for 9 out of 10 classes
 One lecture originated from Pakistan
 Only failure was on the UW-Microsoft Link (which
also brought down UW-Pakistan)


Improved audio (microphone issues)
Participation of students from Pakistan
 Student submissions
 Questions and discussions

Multiple rounds of audio communication
Dec 15, 2008
Microsoft Learning
21




Participants must have incentive for a
distance course
Instructor need to make an effort to create
multisite interaction
Active participants at remote sites help
Time zones and scheduling are major issues
Dec 15, 2008
Microsoft Learning
22
•
Video recorded lectures shown with facilitator
– Original model: lectures stopped by students for
discussion
– Peer tutors
•
•
Developed by Jim Gibbons at Stanford University
Positive results reported in Science [1977]
Dec 15, 2008
Microsoft Learning
23

Introductory programming
 Address community college articulation
 Experiment with alternate approaches to
introductory computing instruction

UW – Beihang Algorithms course
 Offering of UW Algorithms course in China

Digital StudyHall
 Primary education in rural India
Dec 15, 2008
Microsoft Learning
24

Recorded lecture materials
 Generally based on live classes

Class model
 Lecture playback alternating with facilitator led
discussion
 Facilitation models
▪ Gibbons: Peer instruction
▪ Active facilitation
Dec 15, 2008
Microsoft Learning
25
Offer a course in Beijing
based on UW course
 UW instructor could not
teach the course in China
 Scheduling prevented a
live course offering

 1:30 pm in Seattle is 4:30
am in Beijing
 Materials captured from
live class

Tutored Video Instruction
 Slides, talking head, digital
ink
Dec 15, 2008
Microsoft Learning
26

Set up visit





Met with Teaching Assistants
Tested all technology
Trained Teaching Assistants in facilitation
Gave classes to students to demonstrate technology and
TVI
Midterm visit
 Observed classes
 Gave lecture without recorded video


Regular communication with Teaching Assistants
Data collection
Dec 15, 2008
Microsoft Learning
27

Applications displayed
 Webviewer for video
replay
 Classroom Presenter

Teaching Assistants
would show video or
show CP for inking on
slides or classroom
interaction
Dec 15, 2008
Microsoft Learning
28

Offering successful
 Technology, institutional relationship

Cross-cultural issues
 English language materials were comprehensible
 Classroom discussion primarily in Chinese

Facilitation model
 Significant support for facilitators
 Classroom activities successful (and popular)
 Facilitators innovative and reproduced some of the
instruction
 Interactive and informal classroom atmosphere
Dec 15, 2008
Microsoft Learning
29

Lectures delivered in English
 Language exposure consider to be a positive side
effect of the course

Teaching assistants facilitated in English
 But discussions were generally in Chinese


Students reported using lectures outside of class
Instructor observations from site visit
 Chinese students had substantially more English
listening than speaking experience
 Recorded lectures did contain some colloquial usage
and cultural specific references which were lost
Dec 15, 2008
Microsoft Learning
30

Support provided for
facilitators
 Lecture notes
 Activities

Example: facilitators
working through example
from lecture slides
Facilitators invested a
larger effort in preparation
 Studying videos
 Planning how to cover
content


Instructor
Facilitator A
Facilitator B
Facilitator C
Active facilitation
 Worked through lecture
examples
 Led activities
 Asked questions to students
Dec 15, 2008
Microsoft Learning
31

Tablet PC supported
activities
 Student submission model
 Used for every lecture


Technology generally
successful
Considered very positive
by students
 High rate of participation

Provided a structure for
active learning
Dec 15, 2008
Microsoft Learning
32


Contrast to traditional large lecture class
Highly interactive class
 Interaction episodes measured by observation logs
and videos of Beihang classes
 Average of 13 interaction episodes per class, 10 with
students speaking
 UW class averaged about 20 interaction episodes per
equivalent length of time
 Beihang episodes averaged a greater number of
rounds of communication

Class atmosphere was informal
Dec 15, 2008
Microsoft Learning
33
•
•
•
Offering successful
– Technology, institutional relationship
Cross-cultural issues
– English language materials were comprehensible
– Classroom discussion primarily in Chinese
Facilitation model
– Significant support for facilitators
– Classroom activities successful (and popular)
– Facilitators innovative and reproduced some of the
instruction
– Interactive and informal classroom atmosphere
Dec 15, 2008
Microsoft Learning
34
Affiliated Project
Collaboration with
Randy Wang in
Lucknow
• Tutored Video
Instruction for primary
education in rural India
• YouTube + Netflix
•
•
Dec 15, 2008
Microsoft Learning
35




A people’s database
Mediation based pedagogy
Hub and spoke model
Content distribution by DVD
Dec 15, 2008
Microsoft Learning
36
•
•
Operational hubs in Lucknow, Calcutta, Pune,
and Bangladesh
Each hub works with a number of poor village or
slum schools
Dec 15, 2008
Microsoft Learning
37

Classroom study starts July 2009
 Two year study

12 schools in Uttar Pradesh
 12 DSH Classrooms
 12 Traditional Classrooms



Evaluate Learning gains through
pretest/posttest
Classroom observations
Study impact on teachers
Dec 15, 2008
Microsoft Learning
38
•
•
Value of electronic materials in the process of
classroom instruction
Tools for teaching
– Teacher and students drive the process
– Flexible and unpredictable use
•
Importance of high reliability
– And attention to address issues
•
Broader context – interplay of technology
and other issues
Dec 15, 2008
Microsoft Learning
39

Facilitated Video Instruction
 Biggest opportunity to enhance expertise

Synchronous Distance
 Limited applications – but some specific cases of
interest
 Classroom capture

Classroom Computing
 Computer lab and individual devices

Computing Education
 Lack of technical literacy is a limiting factor
Dec 15, 2008
Microsoft Learning
40

Recorded video with a facilitator
 Take advantage of expertise
 Facilitator has very important role

Applicable to a broad range of educational
domains





Primary and secondary
Health
Agriculture
Vocational
College
Dec 15, 2008
Microsoft Learning
41

Applicability for low resource environments
 Low cost
▪ Replay (TV + DVD)
▪ Digital Video + PC based editing
 Locally created content

Key components
 Facilitation methodology
 Support for facilitator
 Program structure
▪ Content creation through delivery

Technology




Video production tool chain
Content distribution
Delivery
Feedback and monitoring
Dec 15, 2008
Microsoft Learning
42

Bandwidth is a significant issue
 Low bandwidth + high cost

May be relevant for institutional outreach
programs
 Arvind Eye Hospital, Tamil Nadu, India
 International medical education


Positive side effect: classroom capture
Risk: Driven by need to justify infrastructure –
not for educational needs
Dec 15, 2008
Microsoft Learning
43

School based computing
 Risk: deployments leading educational goals
 Challenge: maximize benefits of “computers for
schools initiatives”

Models
 Low cost educational devices
▪ Classmate, OLPC, Asus eee, …
 Shared use
▪ Multipoint
 School server
Dec 15, 2008
Microsoft Learning
44


Shortage of trained professionals
Hostile Computing Ecosystem
 Very high virus infection
 Poorly administered machines
 Solutions
▪ Computing practices and maintenance
Dec 15, 2008
Microsoft Learning
45
•
Richard Anderson
– anderson@cs.washington.edu
•
Classroom Presenter
– http://www.cs.washington.edu/education/dl/presenter/
•
Center for Collaborative Technologies at UW
– http://cct.cs.washington.edu/
•
Digital StudyHall
– http://dsh.cs.washington.edu/
Dec 15, 2008
Microsoft Learning
46
Support from Microsoft Research, National Science
Foundation, HP, Ford, UW CSE
• Jay Beavers, Jane Prey, Randy Hinrichs, Chris Moffatt,
Jason Van Eaton, Paul Oka, Steve Wolfman, Ken
Yasuhara, Andrew Whitaker, Ruth Anderson, Craig
Prince, Valentin Razmov, Natalie Linnell, Krista Davis,
Jonathon Su, Sara Su, Peter Davis, Tammy
VanDeGrift, Joe Tront, Alon Halevy, Gaetano Borriello,
Ed Lazowska, Hal Perkins, Susan Eggers, Fred Videon,
Rod Prieto, Oliver Chung, Crystal Hoyer, Beth Simon,
Eitan Feinberg, Julia Schwarz, Jim Fridley, Tom
Hinkley, Ning Li, Jing Li, Luo Jie, Jiangfeng Chen,
Melody Kadenko, Julie Svendsen, Shannon Gillmore
•
Dec 15, 2008
Microsoft Learning
47