attention profiling algorithm for video-based lectures

29
S C I E N C E P A S S I O N T E C H N O L O G Y www.tugraz.at Attention Profiling Algorithm for Video-based Lectures Josef Wachtler, Martin Ebner and Behnam Taraghi ZID - Social Learning - TU Graz HCII 2014

Upload: social-learning

Post on 06-May-2015

524 views

Category:

Technology


0 download

DESCRIPTION

Presentation at HCII 2014, Crete, June 2014

TRANSCRIPT

Page 1: Attention Profiling Algorithm for Video-based Lectures

S C I E N C E P A S S I O N T E C H N O L O G Y

www.tugraz.at

Attention Profiling Algorithmfor Video-based LecturesJosef Wachtler, Martin Ebner and Behnam TaraghiZID - Social Learning - TU Graz

HCII 2014

Page 2: Attention Profiling Algorithm for Video-based Lectures

2

Attention Profiling Algorithm for Video-based Lectures

Content

1. Motivation

2. Implementation of the Algorithm– Operating-Context– Recording Joined Timespans– Calculating the Attention-Level

3. Evaluation

4. Conclusion

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 3: Attention Profiling Algorithm for Video-based Lectures

3

Attention Profiling Algorithm for Video-based Lectures

Graz, University of Technology

Europe, Austria, Graz

http://www.tugraz.at

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 4: Attention Profiling Algorithm for Video-based Lectures

4

Motivation

Students’ Attention

students are confronted with a growing quantity ofinformation

they can handle and process only a limited number ofthese information at the same time

selective attention is the most crucial resource forhuman learning

so it is from high importance to control and analyze it

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 5: Attention Profiling Algorithm for Video-based Lectures

5

Motivation

Interaction and Communication

should be used in many different forms as well as inall possible directions

avoid that learners become tired or annoyed

increase the attention and the contribution

feedback for teachers:

Is it possible for the learners to follow the content?Is the speed appropriate?...

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 6: Attention Profiling Algorithm for Video-based Lectures

6

Implementation of the Algorithm

Overview

the attention profiling algorithm is divided in two parts:

a detailed recording of the joined timespans ofeach single userthe calculation of an attention-level based on thereaction-times to the interactions

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 7: Attention Profiling Algorithm for Video-based Lectures

7

Operating-Context

Web-Application

on-demand video or live-broadcasting

implements the attention profiling algorithm

different methods of interaction:

automatically asked questions and captchasasking questions to the lecturerasking text-based questions to the attendeesmultiple-choice questions at pre-defined positions

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 8: Attention Profiling Algorithm for Video-based Lectures

8

Operating-Context

Interactions during a Video

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 9: Attention Profiling Algorithm for Video-based Lectures

9

Recording Joined Timespans

Functionalities

for each attendee it is possible to say

at which time he/shewatched which part of the video

calculating statistical values

the shortest or the longest joined timespanthe average length of the joined timespans...

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 10: Attention Profiling Algorithm for Video-based Lectures

10

Recording Joined Timespans

Models

the JoinedUser -model connects an user to an event

the History -model represents a joined timespan withboth, absolute and relative timestamps

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 11: Attention Profiling Algorithm for Video-based Lectures

11

Calculating the Attention-Level

Functionalities

calculation of an attention-level which is based on thereaction-times of the attendees to the interactions

1. logging the reaction-times

2. calculating the attention-level

maxim: if the attendee reacts slower theattention-level decreasesresult ranges from 0% to 100%

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 12: Attention Profiling Algorithm for Video-based Lectures

12

Calculating the Attention-Level

Logging the Reaction-Times

the Interaction-model

with its concrete sub-class models for eachpossible receiver of an interactionconnects an interaction to an user

the CallHistory-model

logs every occurrence of an interactionin absolute and relative timestampsthe difference between the real start time and theresponse time is equal to the reaction-time

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 13: Attention Profiling Algorithm for Video-based Lectures

13

Calculating the Attention-Level

Logging the Reaction-Times

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 14: Attention Profiling Algorithm for Video-based Lectures

14

Calculating the Attention-Level

Overview

the calculation is split in three rounds:

1. calculation of an attention-level based on thereaction-times for every call of an interaction (I)

2. grouping them to attention-levels (AL) of eachinteraction-methods (IM)

3. generalizing to an attention-level of a joined timespan

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 15: Attention Profiling Algorithm for Video-based Lectures

15

Calculating the Attention-Level

Round 1

the calculation has two parameters

1. SUCCESS UNTIL states the time until anattention-level of 100% could be reached

2. FAILED AFTER indicates after whichreaction-time an attention-level of 0% will beassumed

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 16: Attention Profiling Algorithm for Video-based Lectures

16

Calculating the Attention-Level

Round 1

f (tij) represents the attention-level of the j-thinteraction of the i-th interaction-method

tij is the corresponding reaction-time

f (tij ) =

100 if tij ≤ SUCCESS UNTIL0 if tij > FAILED AFTERg(tij ) else

(1)

Where g(tij) is

g(tij ) = 100 −(

tij − SUCCESS UNTILFAILED AFTER − SUCCESS UNTIL

∗ 100)

(2)

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 17: Attention Profiling Algorithm for Video-based Lectures

17

Calculating the Attention-Level

Round 2

ai calculates the attention-level of the i-thinteraction-method by forming the mean

mi is the number of its interactions

ai =

mi∑j=0

f (tij)

mi(3)

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 18: Attention Profiling Algorithm for Video-based Lectures

18

Calculating the Attention-Level

Round 3

takes the attention-level of each interaction-method(ai) and again forms the mean over them

n is the number of interaction-methods

attention =

n∑i=0

ai

n(4)

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 19: Attention Profiling Algorithm for Video-based Lectures

19

Evaluation

Overview

three goals

1. gaining suitable parameters to force the algorithmto deliver realistic values

2. comparing the results of the algorithm with thefeedback of the attendees to implement adoptions

3. evaluating the effects of the adoptions

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 20: Attention Profiling Algorithm for Video-based Lectures

20

Evaluation

Gaining Suitable Parameters

live-broadcasting of the lecture Societal Aspects ofInformation Technology

analyzing recorded reaction-times of the interactions

the average reaction-time is calculated to place theparameters around this point

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 21: Attention Profiling Algorithm for Video-based Lectures

21

Evaluation

Compare Results with Feedback

live-broadcasting of the lecture Introduction toStructured Programming

complete number of attendees vs. active ones

active: watched ≥ 75% and attention-level ≥ 50%

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 22: Attention Profiling Algorithm for Video-based Lectures

22

Evaluation

Results

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 23: Attention Profiling Algorithm for Video-based Lectures

23

Evaluation

Feedback

attendees felt uncomfortable with their attention-level- they assumed a much higher one

impossible to answer faster because the live-streamdoes not stop if an interaction occurs

the number of interactions should not be very high

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 24: Attention Profiling Algorithm for Video-based Lectures

24

Evaluation

Adoptions

the video pauses if an interaction occurs

lecturers asked to pause his/her presentation at theoccurrence of an interaction at a live broadcasting

the number of interactions is lowered to a maximumof three interactions in a period of ten minutes

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 25: Attention Profiling Algorithm for Video-based Lectures

25

Evaluation

Testing Adoptions

8 videos of the lecture Learning in the Net: Frompossible and feasible things

complete number of attendees vs. active ones to testthe adoptions

active: watched ≥ 75% and attention-level ≥ 50%

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 26: Attention Profiling Algorithm for Video-based Lectures

26

Evaluation

Results

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 27: Attention Profiling Algorithm for Video-based Lectures

27

Evaluation

Discussion

the parameters for the calculation of theattention-level are highly sensitive

the accuracy depends on many different factors (e.g.difficulty of the questions, the content of the video, ...)

the timespan between the interactions should not beto small

the two parts of the attention profiling algorithm areonly powerful in combination

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 28: Attention Profiling Algorithm for Video-based Lectures

28

Conclusion

Conclusion

attention is the most crucial resource in humanlearning

attention profiling algorithm with to parts

recording of the joined timespanscalculation of an attention-level

delivers realistic values after some adoptions

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014

Page 29: Attention Profiling Algorithm for Video-based Lectures

29 Thank you ...

... for your attention!Questions?

Josef Wachtler, [email protected] Ebner, [email protected]

ZID – “Social Learning”Graz, University of TechnologyMunzgrabenstraße 35A, A-8010 Grazhttp://elearningblog.tugraz.at

Josef Wachtler, Martin Ebner and Behnam Taraghi, ZID - Social Learning - TU GrazHCII 2014