Konstantin Filtschew presents examples of how eLearning systems can provide educational assistance through interactive exercises in various subjects like vocabulary, history, math, and foreign languages. Natural language processing techniques like part-of-speech tagging and parsing can help generate automatic questions and analyze student answers, though human oversight is still needed. Mobile devices open opportunities for learners to make use of spare time for educational podcasts, tutorials, and customized information on the go.
2. About me
Name: Konstantin Filtschew
Interested in:
Innovation
eLearning Systems
Security in computer systems
Software design
New challenges
...
2
4. Do we need ”eLearning”?
Wissensgesellschaft
General Knowledge
Specific Knowdledge
Transfer of knowledge to the next generation
cp -r / remotebrain: # maybe in future
We can't stop learning!
4
5. Definitions of eLearning
Internet-enabled learning that encompasses training, education, just-in-
time information, and communication.
(http://www.eng.wayne.edu/page.php?id=1263)
Any learning supported by digital means.
(http://azelearning.org/glossary/3)
E-learning (or electronic learning or eLearning) encompasses forms of
technology-enhanced learning (TEL) or very specific types of TEL such
as online or Web-based learning.
(http://en.wikipedia.org/wiki/E-learning)
eLearning kann verstanden werden als ein Lernprozess, der durch
Informations- und Kommunikationstechnologie unterstützt wird
(http://www.uni-hildesheim.de/de/9808.htm)
5
6. Assistance through eLearning
People must try to get better user experience
People must do exercises for better knowledge
and understanding
Some examples
Vocabulary
History
Math
Foreign languages
6
7. Vocabulary Assistance
Vocabulary trainer
Does the user really know the word?
Interactive trainer
Enter vocabulary once
(semi automatic possible)
Shows vocabulary randomly
Shows/checks correct answer
instantly
Can analyze time for answers
7
Vocabulary Frequency
8. History eLearning example (1)
We shouldn't try to replace books and paper
Add interactive media (sound/video)
Ineractive questions to text
Multiple choice
Plain text answers (keyword NLP)
●
+
Who was Cristopher Columbus?
In which year he started his travel?
+ +
●
● Was it India he found on his first travel?
8
● ...
9. History eLearning example (2)
We have some problems to solve first!
Many different books used
People are lazy to search for more information
Need to reference the learners book, text and part
Need help to create tasks (NLP can help)
Questions and multiple choice generation
Help: Google books scan project 9
10. Math eLearning
Some examples:
Gehirntraining
Math quizzes:
X + 11 = 23
126 / X = 42
sqrt(169) = X
Motivation:
Points (Challenge)
Success experience
10
Not the same exercises
11. Foreign Language
Common difficulties in languages: Prepositions
Both Parties had much to offer ________ a time of growth in the region.
(a) in (b) at (c) about
Part of Speech Tagger or Parser
Can recognize prepositions
Corpora
Help identify wrong prepositions:
Check whether this combination is really wrong
Common with left word of prepositions but not
with right and vice versa 11
13. Interactive Tutor (2)
Natural Language Processing (NLP)
Questions by tutor
Answers in plain text
Tutor helps to ...
find answer
Helps to remember all important points
Problems:
Still need to reference the lecture
13
Not very ”human” in handling (no emotions)
14. Natural Language Processing
NLP is a great help:
Analyze Text and extract important information
Can generate questions (almost automatic)
Can analyze answers
Need:
Corpora: digital texts
Part of Speech Tagger and Parser
A lot of computer power for given tasks
Human control for results 14
15. NLP: Corpora
We have a lot of digital text:
Wikipedia (and other wikis)
Blogs
Google book scan
Problems:
Common Speech (Umgangssrpache)
Errors
(Internet) slang → :) :/ ;) cu
Ambiguity (Ambiguitäten) 15
16. Part of Speech Tagger and Parser
Part of Speech (POS) Tagger (Wortart Tagger)
Analyzes only words
Not so precise (about 80-90%)
Faster than parser
Parser
Additionally to POS: sentence structure
More precise (up to 96%)
Sentence tree (word relations)
16
18. Sentence Tree
From Phineas Q. Phlogiston, “Cartoon Theories of Linguistics, Part ж—The Trouble with NLP“,
Speculative 18
Grammarian, CLIII(4), March 2008
19. Human control
Let people learn errors is a very bad learning
experience
Wrong answers can be still right
Computer can't decide about uncommon tasks /
questions / answers
Advantage:
Automatic generation of tasks
Teacher has only to check the
written tasks 19
20. Digital Books / Newspapers
Add interactive tasks to common media
Digital Books:
We can add additional information (Video/Audio)
Add exercises and quizzes
Digital Newspapers:
Schools can use for lessons (teacher selects)
Teacher can create semi automatic tasks
Advantage for newspapers: children learn to read
newspapers 20
21. Mobile learn experience (1)
We have:
Powerful mobile phones: IPhone / Android / …
Book/Newspaper readers
(not yet open for extensions)
We have to travel
Home → Work → Home
Business travels
Time slots for education
21
22. Mobile learn experience (2)
Busy people use every free minute
Maybe you can't work, but you can learn
Allready in use as eLearning:
Podcasts
Video and audio tutorials
Already filtered information
(Read it later add-on)
22
23. Conclusion
We have a lot to learn
We have to use our free time slots
We allready use eLearning
eLearning is at the very beginning
Politics say: Wissensgesellschaft
23
24. Thank you for your attention
Questions?
konstantin@filtschew.de
http://konstantin.filtschew.de/blog/
http://twitter.com/Fa11enAngel
24
25. Sources
http://upload.wikimedia.org/wikipedia/commons/6/6e/Latin_dictionary.jpg (GPL) //Books
http://media.photobucket.com/image/ship/MODELSHIPCONSTRUCTION/17Th%20Century%20Man-of-War/NewBritishShip.jpg
//Ship
http://en.wikipedia.org/wiki/File:Ridolfo_Ghirlandaio_Columbus.jpg // Christopher Columbus
http://web.airgamer.de/fileadmin/airgamer/images/spiele/01-gehirntraining/ss_gehirntraining_01.png //Gehirntraining
http://upload.wikimedia.org/wikipedia/commons/3/38/Gregor_Reisch%2C_Margarita_Philosophica%2C_1508_%281230x1615%29.png
// Gregor Reisch, Margarita Philosophica
http://openclipart.org/people/StefanvonHalenbach/StefanvonHalenbach_Teacher_L_mpel.png // Teacher
From Phineas Q. Phlogiston, “Cartoon Theories of Linguistics, Part ж—The Trouble with NLP“, Speculative Grammarian,
CLIII(4), March 2008 // Pretty little girl
25