|
|
|
Michel Plu, Christine Chardenon, Laurent Maupeu |
|
France Telecom R&D Lannion FRANCE |
|
|
|
|
|
|
|
World wide networks give access to thousands of
lively multilingual audiovisual
resources |
|
Digital technology facilitates the management
and use of accessible resources |
|
|
|
A real opportunity to teach foreign languages
with more diverse, interesting and up-to-date audiovisual resources |
|
The power of images |
|
Let pupils choose the contents they really
appreciate |
|
Raise
motivation, a key to education |
|
|
|
|
|
Using videos sometimes raises copyright and
legal issues |
|
More and more materials become free of use for education purposes |
|
Searching videos is time-consuming |
|
Indexation is hard , inaccurate or expensive |
|
Choosing videos is longer than choosing text |
|
Audiovisual resources are often harder to
understand than written text |
|
|
|
Lots of fun, but what do students really learn ? |
|
Formal knowledge has also to be taught with more
« conventional » activities |
|
|
|
|
|
An experimental online service available via the
web to help teachers to |
|
|
|
Take advantage of the multiple resources available to build original, customized
multimedia sessions |
|
Find the appropriate resources according to the
skill levels of their pupils and their pedagogical objectives |
|
Build pedagogical materials that pupils can
understand |
|
Motivate their pupils with lively and funny
activities |
|
Integrate conventional activities like grammar
exercises with multimedia sessions |
|
|
|
Using broadcasted closed-captioned videos |
|
|
|
|
|
|
Video search engine with pedagogical criteria |
|
According to |
|
Speed of speech |
|
Level of vocabulary used |
|
Syntactic patterns or grammar forms to be
learned |
|
Combined with classical keyword based search |
|
Help for navigating into the video |
|
To have
the current spoken sentence being repeated or grammar patterns being
searched |
|
Markers can be set to let pupils go to these
sentences |
|
Help for understanding vocabulary |
|
With controlled access to dictionary definition |
|
Tools to generate “gap-fill” exercises from the
closed-caption text of the videos |
|
|
|
|
|
|
|
|
|
|
|
|
Record
only closed-captioned broadcasted contents with a set top box |
|
|
|
Split them up into short clips |
|
|
|
Extract closed-caption text from the videos |
|
Generate ascii file with time code information |
|
One day closed caption might be automatically
generated with “speech to text technology” |
|
|
|
Encode broadcasted videos in order to stream
them into a server |
|
|
|
|
|
|
|
The closed caption text is parsed and processed
with a Natural Language Processing Tool |
|
|
|
Paragraphs, sentences and words are identified |
|
Each word is tagged with |
|
Its lemma , part-of speech, and morpho-syntactic
properties: |
|
like
number, person, form, tense, mode for verbs, positive, negative superlative
for adjectives … |
|
Time codes are tagged like specific keywords |
|
|
|
Result is in XML |
|
|
|
|
|
|
|
Parse XML file to compute pedagogical indexes |
|
Difficulty to understand the video |
|
Speed of speech |
|
based on
the average number of word syllables pronounced between two time codes |
|
Vocabulary level |
|
Percentage of spoken
words in the video in list of words corresponding to skill level in the
language |
|
Grammar profile |
|
number
of times a specific grammar form or syntactic pattern is found in the NLP
result |
|
These indexes are used by the teacher to select
videos |
|
adapted to their pupils’ skills or their
pedagogical objectives |
|
|
|
|
|
|
|
|
|
Integration of synchronized closed caption text
and streamed video |
|
Generation of a
SMIL and RealText™ formatted file |
|
The RealText™ closed caption is augmented with : |
|
A navigation arrow to rewind from one sentence
to another |
|
Each
time a new sentence is detected the time code associated to the navigation
arrow is modified |
|
Links to go to a specific sentence with a
particular grammar form |
|
Links with difficult words to textual
definitions from an online dictionary |
|
Each
entry has its own URL |
|
Use list
of commonly known words |
|
Use NLP
result to find the right entry in the dictionary |
|
|
|
SMIL files of selected videos are customized
online, retrieved and stored onto the school server |
|
|
|
|
|
|
|
For each closed-caption text several HTML
“gap-fill” exercises are generated |
|
Use of pattern substitution rules to generate an
HTML form |
|
replace
specific grammar or syntactic patterns with an input field |
|
Generation of indication to fill the gap |
|
lemma
form of the replaced words and morpho-syntactic properties found |
|
Evaluation of the answers by a script on the
http server |
|
comparing input text with original form found in the text |
|
|
|
Several kinds of exercises can be generated |
|
Verbs, adjectives with comparative or
superlative form, prepositions |
|
|
|
Exercises can be retrieved and edited by
teachers and stored onto the school server |
|
|
|
|
|
|
|
12 teachers |
|
College teacher , and teachers for adults in
companies |
|
Global satisfaction and very promising but: |
|
Distinguish skill levels of vocabulary and
targeted audience of the video |
|
Example
: some cartoons are easy to understand for adult learner but unfortunately
are targeted to young children |
|
Other criteria : Number of people
speaking,background noise, accent … |
|
Some contents are more or less adapted |
|
Cartoons
for the young |
|
Music
depends on pupils’ tastes |
|
News is
rapidly out of date |
|
Particularly adapted to adult training sessions |
|
With appropriate materials |
|
|
|
|
|
Networks can provide diverse audiovisual
contents to learn foreign languages |
|
Digital technology can help teachers to use
these new contents efficiently |
|
Closed-captioning is not only important for
people with hard hearing difficulties, but can also provide support for
education abroad |
|
Online Multimedia development will certainly
raise new ways of teaching foreign languages |
|
It’s only the beginning |
|
Adding modernity to modern language teaching |
|