Training at Aeroflot

1
Training at Aeroflot
INRAL Ltd.

• Using a virtual environment for aviation English
  training

2
CALL system for preparation to the ICAO English
tests (implemented in 2005 in Aeroflot Training
Center)
yes
Qualification exams
Periodical consultations
ICAO level 4
Self-training On-line courses, collection of
training materials on CD
no

Language skills evaluation (Annual testing
defining ICAO level)
Individual course correction
Adaptive individual course designing, planning
consultations and tests
2
3
The aim to increase the efficiency of
communication skills training in aviation English
learning system.

• Two lines of development
• creating or choosing effective virtual
  environment for organizing virtual consultations
  and classroom like training with real teacher
• developing new software for learning English
  through training dialogues ( using of chatbots
  and virtual agents) for self-paced training,
  replacing real teacher with virtual assistant

4
First line of development virtual environment
for aviation English consultations and training

• Main tasks
• To develop cost-effective on-line solution which
  will allow to imitate the classroom group
  training and consultations
• 2. To adopt the existing aviation English
  learning methodology for training in virtual
  environment
• 3. To modify the existing computer training
  courses and materials for
• using in virtual environment

The solution to build virtual environment based
on virtual world technology.
5
Training in virtual environment the dialog
between teacher and trainee or trainees.

• Training dialog could include such objects as
• Presentations
  - Audio files
• - Video files
  - various exercises and tests.

6
General system scheme
Virtual Environment
Administrator Interface
Student Interface
OpenSimulator Server
Statistics gathering service
Internal Web browser
Teacher Interface
Learning Management System (LMS)
Legend
Assets
Statistics DB
HTTP protocol
Units (Lessons) DB
DB links
Database module
7
Second line of development using of chabots for
self-paced aviation English training, replacing
real teacher with virtual assistant.
A chatterbot (or chatbot) is a type of
conversational agent, a computer program designed
to simulate an intelligent conversation with one
or more human users via auditory or textual
methods. (www.wikipedia.org)
A.L.I.C.E (Artificial Linguistic Internet
Computer Entity) - the most famous chatbot
developed by Richard Wallace. The program uses
XML Schema called AIML (Artificial Intelligence
Markup Language) for specifying the heuristic
conversation rules. The basic knowledge unit
(category) in AIML consists of an input
question (pattern), an output answer
(template), and an optional context. The
dialogue is based on algorithms for automatic
detection of patterns in the dialogue data.
8
Virtual Assistant as a community of chatbots with
special competenties.
Latent semantic analysis (LSA) is mathematical
method for computer modeling and simulation of
the meaning of words and passages by analysis of
representative corpora of natural text. LSA
closely approximates many aspects of human
language learning and understanding.
User Input
Spellshecker
LSA (Latent Semantic Analysis)
Grammar Checker (Parser)
Grammar question
Grammar error
Radiotelephony chatbot
Aviation English chatbot
Generic chatbot
Grammar chatbot
Answer
9
Radiotelephony chatbot
General control block
XML scenario
User input
AIML set 1
AIML sets
AIML set N
ICAO doc.4444 (Rules of the AIR and Air Traffic
Services, Part X Praseologies) Example Circumstan
ces Level changes, reports and
rates. Phraseologies a) CLIMB (or DESCEND)
followed as necessary by
TO (level)
TO REACH (level) AT (or BY)
(time or significant point)
REPORT LEAVING (or REACHING, or PASSING)
(level) REPORT
PASSING ODD (or EVEN) LEVELS
STOP CLIMB (or DESCENT) AT (level)
REQUEST LEVEL CHANGE FROM
(name of unit) AT (time or significant point)
and so on.
10
General system scheme
Virtual assistant
Set of expert chatbots
Aviation English chatbot
Virtual Environment
User question analyzer
Administrator Interface
Radiotelephony chatbot
English Grammar chatbot
Pattern processing module
Student Interface
OpenSimulator Server
Pseudo Client (Avatar)
Generic chatbot
Statistics gathering service
General Control module
Internal Web browser
Teacher Interface
Learning Management System (LMS)
Legend
Statistics DB
HTTP protocol
Units (Lessons) DB
Assets DB
user question-chatbot answer loop
Database module
Direct avatar control
DB links
11
Unit (lesson) structure (XML file)
Element Attributes Description
Unit Name Top level element for each unit
Presentation id-the programmatic name of presentation Specifies the set of frames
Frame idf-the programmatic name of frame Specifies one step of presentation
Asset ida the name of asset in database Specifies the media object for given frame
Text Specifies the explanation text for given frame
Action id the programmatic name of action Specifies the action of avatar in virtual environment for given frame
Loop Specifies that actions should be repeated
Ask_me_dialog id_pres the id of presentation time the maximum time of silence Specifies the dialog in which avatar answer to the student questions, optionally the questions are connected with presentation.
Quest_dialog id_pres the id of presentation time the maximum time of silence Specifies the dialog in which avatar asks the student
Start_text Specifies the words of avatar at the beginning of the dialog
Question Q_text- the text of question H_text- the help text, which avatar tells if the student silence exceeds the specified time Specifies the main question in Quest_dialog
12
Presentation
13
Quest_dialog
14
Using of video fragment
15
Scheme of working prototype