Skip to content

Instantly share code, notes, and snippets.

@Harrison-M
Created July 13, 2013 17:25
Show Gist options
  • Save Harrison-M/5991428 to your computer and use it in GitHub Desktop.
Save Harrison-M/5991428 to your computer and use it in GitHub Desktop.
TX Dev Fest A&M Research Projects notes
A&M Research Projects
David Turner, Manoj Prasad, Dr. Tracy Hammond
Platforms: Android, Google App Engine
API's: Google Maps
Mechanix: Free body diagram tutoring program
Predictive labeling system
Checklist and examination system guides user through process of creating a free body diagram
Tutorial video on YouTube
Currently at A&M in classes
Has been shown to improve learning
Does not scale well (fails at ~30 users)
Only supports one domain
Does not allow for easy expansion
Only works at A&M
Did work at one other school, but harrowing moments
Better: Course Sketch
Presentation on YouTube
Sketch Recognition Lab developing MOOC platform that uses sketch recognition for automatic grading
Currently supports Physics, Chemistry, and Kanji
Web app
Learning outside classroom
Challenges
Provide personalized feedback
Internationalization
Low interaction overhead/natural interaction
Education for whole world
Currently
Provides feedback
Supports 3 domains
has courses
Runs on App Engine
Backend goals
Send and Receive data quickly
Remain responsive under stress
Maintain communication during sketch while conducting multiple simultaneous recognition processes
Tolerate intermittent network disconnect
Interface design
Used App Engine to make quick iterations
Can scale quickly without load concerns
Cheap
Easy API
Interaction process
Begin stroke - starts recognition
Stroke is sent back
Model is created/updated
Recognition is sent out to other App engines
Different results are compiled
User is supplied with updated model
Flexible
App engine allows flexible network
Ability to outsource recognition to external developers
Makes network easier to distribute
Allows easy extensions & growth
Connecting externally
External developers host their own app engines to do more specific domains (courses)
If we can't answer a request, send the request through HTTP
Future plans
Create API for recognition and checking answers
Answer/domain creation
Enhanced interaction for domains
More specific object detection
Ability to connect external developers
Animal Games
Ad-hoc multiplayer games
Smartphones/Tablet
Traditional console environment does not allow for private information in games
Demo on YouTube
Campus Compass
TAMU campus specific mapping service
iOS/Android/Web
Locate buildings on campus
Route to buildings
Demo video on YouTube
World of Workout
Exergaming on smartphones
Mapping physical attributes to avatar in game
Use accelerometer and GPS
Workout amounts determine stats of character
Good Bytes
Restaurant sentiment analysis
Dr. Baker - language analysis research
Demo video on YouTube
Kamistry
Teaching Chemistry
Augmented reality + gestural interface
Kinect + Android/iOS
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment