Learning Traces, Learning Clocks and Artificial Observers

1
Learning Traces, Learning Clocks and Artificial
Observers

• Alvarez-González, Luis A.
• Universidad Austral de Chile.
• November, 2007

2
Content

• Introduction
• Learning System Model
• Learning Traces
• Learning Clocks
• Artificial Observer
• Conclusions

3
Introduction

• CSCL, emerges as a discipline 1989.
• The concepts of Learning Objects was developed at
  the beginning of 2000.
• The standard IMS LD of Learning Design was
  published at the beginning of 2003.
• The first learning design management system
  appears in 2002. And good example is LAMS.

4
Introduction

• The learning design managements systems are a
  good set of tools. However, are not enough to
  know the behavior and aid to students group,
  specially in large groups.
• Its important to know the learning trace (Unit
  of Learning completed by students) of each one
  and the learning trace of the group as a whole,
  in an asynchronous environment.
• Where the learning trace is a details

5
Introduction

• Finally, A good learning depends mainly on a good
  design and the teacher.
• The learning technologies can only
• aid to teachers !
• So, in this proposal a good learning design is
  assumed.

6
Introduction

• How to do that ?
• The task could be complex and we need a
  theoretical model to help us.

7
Learning System Model
8
Learning System Model
Learning System is a team of collaborators using
a learning design management system.

• Collaborators are
• Students. Following a learning design.
• Artificial observer. Looking the learning
  process.
• Tutors. Guiding the learning process.
• All of they interconnected by a network
• The Learning Designs Managements Systems will be
  on a server.

9
Learning System Model

• An learning design could be considered as
  sequence of learning activities ai
• where every activity can be one learning object
  or several optional learning objects.
• And each student following the sequence

10
Learning System Model
a1
a2
a3
Note that the four types of LO are in the figure
1. Instruction 2. Colaboration 3. Simulation
and 4. Evaluation
a4
Luis Alvarez
Felipe Zapatero
Pedro Pérez
a5
a6
Juan González
a7
a8
a9
11
Learning Traces
12
Learning Traces

• In an asynchronous system there exist no
  constrains on the learning relative speed.
  However, is important to help the slow
  students.
• If we know the sequence of learning activities
  completed by an student (learning trace) is
  possible to know the group learning trace.

13
Learning Traces
a1
a2
a3
a4
a5
a6
a7
a8
14
Learning Traces

• Let, learning trace for the student i is the
  ordered set of activities completed by student i
• tiai1 ai2 .
• And learning trace of all students as a set
• T t1? t2? ?tn.

15
Learning Traces
a1
a2
a3
a4
a5
a6
a7
a8
t(Pedro Perez) a11, a21, a31 , a41, a51
t(Luis Alvarez) a12, a22, a32 T
t(Pedro Perez) ? t(Luis Alvarez) ? t(Felipe
Zapatero) ? t(Juan Gonzalez)
16
Learning Clocks
17
Learning Clocks

• Today, usually the students do many things at the
  same time.
• Every students has his/her own learning speed
  and then own learning clock.
• The learning speed is not constant. Could be
  fast at the beginning a slow at the end.
  Moreover, depend of external factors.

18
Learning Clocks

• A single natural number is sufficient to
  represent the set ti(a)
• Learning traces can be represented as 2-
  dimensional vector ti (uidi , lcki )
• uidi unique user identifier, of the student i
• lcki learner clock of the student i.
• The entire learning trace can be represented by
  an 2 x n dimensional matrix.
• T t1T, t2T,.., tnT

19
Learning Clocks

• Each learning clock will be incremented with
  every learning activity carried out for
  corresponding student.
• So, for example the learning clocks are
• For Pedro Perez, t1 (50,5)
• For Luis Alvarez, t2 (53,3)
• For Felipe Zapatero, t3 (51,5) and
• For Juan Gonzalez t4 (49,6)
• NOTE Luis Alvarez need help!

20
Learning Clocks
(50,4)
(50,2)
(50,3)
(50,5)
(50,1)
(53,2)
(53,3)
(53,1)
(51,4)
(51,5)
(51,2)
(51,3)
(51,1)
(49,2)
(49,3)
(49,4)
(49,5)
(49,6)
(49,1)
21
Learning Clocks

• And the Group Trace will be

T t1T, t2T, t3T, t4T
22
Artificial Observer
23
Artificial Observer

• Looking for the process.
• Measuring of knowledge of each one and the
  group
• Quantity of learning activities carried out for
  each one and the all group.
• Notifying to Tutor the progress of the group and
  the each student in the learning design.

24
Artificial Observer
Artificial Observer
Student 1
T t1T, t2T, t3T, t4T
t1 (50,5)
T
Student 2
t2 (53,3)
Learningware on a server
Student 3
t3 (51,5)
Tutor
Student 4
t4 (49,6)
25
Artificial Observer

• The Artificial Observer periodically will do a
  multicast to the lesson where a lesson is a
  learning design running in a LDMS and each
  learning process will reply with the learning
  clock and eventually others parameter as score
  for each learning event.
• And it send to tutor by LMS the
• Learning trace for each one and for the all
  group.
• The LMS can calculate parameter as speed learning
  average, progress average, build learning curves,
  etc. With this information the tutor can
  re-design the learning activity.
• Will be one observer for each lesson.

26
Conclusions
27
Conclusions

• A model of asynchronous communication for
  learning groups have been developed.
• Knowledge from other areas can be used in CSCL,
  in this paper we use concepts and definitions
  adapted from process communication in distributed
  systems.

28
Future Work.

• We must develop
• Artificial observers.
• And so on..

29
Thank you for your contributions