Download - Interest Management
![Page 2: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/2.jpg)
2
Overview of AugCog
“AugCog is a DARPA project to extend, by an order of magnitude or more, the information management capacity of the human-computer warfighting integral by developing and demonstrating quantifiable enhancements to human cognitive ability in diverse, stressful, operational environments.”
-LCDR Dylan Schmorrow, PhDAugCog Program Manager
![Page 3: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/3.jpg)
3
AugCog Specifics
AugCog is designed to bring together researchers in a wide variety of disciplines
Goal is to bring more of the brain to the task
A Brain on Today’s HCIA Brain on Augmented Cognition
“I am Denny’s Brain”
![Page 4: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/4.jpg)
4
MOVES Institute Mission
Research, application and education in the grand challenges of modeling, virtual environments and simulation. 3D Visual Simulation Networked Virtual Environments Computer-Generated Autonomy Human Performance Engineering Technologies for Immersion Defense and Entertainment Collaboration Evolving Operational Modeling
![Page 5: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/5.jpg)
5
Outline
The Context MachineLiterary InspirationsGoals and Objectives of the Context MachineTechnical ApproachShort Term ImplementationPayoff and Impact
![Page 6: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/6.jpg)
6
Literary Inspirations
Diamond Age, Neal Stephenson Futuristic novel of a young girl who
discovers an interactive book which plays scenes demonstrating proper behavior of a young lady
She is able to drastically change her life by adhering to lessons learned in the book
![Page 7: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/7.jpg)
7
The Context Machine
So we came up with the idea of a context machine that determines the “context” for the individual that possesses the machine and provides advice/guidance to the bearer.
![Page 8: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/8.jpg)
8
What this Means in Terms of Interest Management
The context machine is envisioned to add to the user’s interest management in several ways:
By informing the user of things he has not yet observed. By informing and filtering for the user things he doesn’t have the time
to observe. By explaining things the user can’t understand By providing an intelligent interface for the user
![Page 9: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/9.jpg)
9
Where Does the Context Machine Fit in AugCog?
LocationLocation
ActionsActions
PeoplePeople
ObjectsObjects
Outer World Outer World AbstractionAbstraction
Determine Determine Context Context
and COAand COA
Output to Output to UserUser
User’s User’s CommandsCommands
Goal 1Goal 1
Goal 4Goal 4 Goal nGoal n
Goal 2Goal 2Goal 3Goal 3
PreferencesPreferences
Visual Visual DisplayDisplay
Aural Aural DisplayDisplay
![Page 10: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/10.jpg)
10
Goals and Objectives of the Context Machine
Develop a robust, dynamic representation of knowledge which is sufficient to denote context
Create a method to accurately transform symbolic sensory data from the environment into a data structure which can be used to determine context
From the current context, determine a course of action that best meets the user’s goals
Accommodate shifting and divergent goals
![Page 11: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/11.jpg)
11
Developing a Representation of Knowledge to Denote Context
Current research in context sensing is being done largely in the area of wearable computers and user modeling User Modeling
Most user modeling applications in this area deal with anticipating user’s needs and presenting useful information - very little in using context to advise user
User modeling normally assumes a static model of the user Wearable Computing
Most current applications in wearables which use context limit scope to location only
![Page 12: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/12.jpg)
12
Current Definitions of Context
“Context is any information that can be used to characterize the situation of an entity. An entity is a person, place, or object that is considered relevant to the interaction between a user and an application, including the user and the application themselves.” - Dey, 99
“Context is usually understood as information that the system can sense and process to facilitate human-computer interaction.” - Salber, 98
“Context is the set of environmental states and settings that either determines an application’s behavior or in which an application event occurs and is interesting to the user.” - Chen, 00
“We define context to be any information that can be used to characterize the situation of an entity, where an entity can be a person, place, or physical or computational object.” - Brown, 98
![Page 13: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/13.jpg)
13
Most Current Definitions of Context Are Lacking an Important Element
Context must be a function of both the environment and the user
Context changes constantly and is difficult to sense from overt information alone Demands a dynamic architecture Demands a learning mechanism
![Page 14: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/14.jpg)
14
Transforming Symbolic Data into a Contextual Data Structure
Transform a stream of symbols from the sensory component into a useful data structure that can be used to capture the essence of “context”
Must be dynamic, since what data is needed depends heavily upon the current context
Data structure must have memory Gives object/knowledge permanence A relationship learned today is probably still valid
tomorrow
![Page 15: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/15.jpg)
15
Determine a Course of Action to Meet the User’s Goals
Must be based upon our definition of “context” which includes: The surrounding environment
Objects, people, location, etc. The user
Goals, urgency, priorities, etc.Combining these two gives the user’s context
![Page 16: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/16.jpg)
16
Accommodate Shifting and Divergent Goals
The user’s goals can shift quickly Discard current goal in favor of another Lower the priority of the current goal in favor
of anotherThe user’s goals can often conflict with one
another Should be able to sense the conflict and offer
planning advice towards resolution
![Page 17: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/17.jpg)
17
Technical ApproachWe are pursuing a concurrent two-tracked
approach: Track 1 involves performing literature search of
current research into context and developing a taxonomy and hypothesis regarding what is necessary and sufficient to create context-aware applications
Track 2 involves creating applications which are context-aware and testing our hypotheses These applications begin as context systems which are
simple and only work in fairly specific domains and increase in complexity and generality
![Page 18: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/18.jpg)
18
Technical ApproachAgent-based machine learning approach:
Will use genetic algorithms to learn both the user and the environment Mainly want their ability to adapt to new environments that
don’t directly match any previously encountered
Currently anticipate building two different types of agents Symbolic constructor agents (SCA) are used to build
a “topography” in an inner world which represents the outer world
Symbolic reactive agents (SRA) explore this inner world to search the topography and make judgments, inferences and recommendations
![Page 19: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/19.jpg)
19
Agent Architecture
Agent (Actor)
SCASCASCASCASCASCA
Symbolic Inner
Environment SRASRASRASRA
SRA
Environment
Agent ActionsSensed Input
![Page 20: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/20.jpg)
20
Short Term Implementation
Intend to demonstrate rudimentary context sensing capability
Creating an application which can determine the context of a situation and react appropriately
Chosen Environment:Military unit on patrol
![Page 21: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/21.jpg)
21
Why a Military Patrol?
Significantly reduced environmentEasier to determine context
Goals Shift RapidlyDoing this for DARPA
Demonstrates military application
![Page 22: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/22.jpg)
22
![Page 23: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/23.jpg)
23
![Page 24: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/24.jpg)
24
![Page 25: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/25.jpg)
25
![Page 26: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/26.jpg)
26
Payoff/Impact
A mechanism for accepting symbolic information from a sensory stream and inserting it into a robust dynamic architecture for representing context
Capable of learning new and novel situations while in use
Monitors user’s goals and assists where appropriate
Adaptable to wide variety of users
![Page 27: Interest Management](https://reader035.vdocument.in/reader035/viewer/2022062520/56815da5550346895dcbd259/html5/thumbnails/27.jpg)
27
Contact Information
Michael ZydaThe MOVES InstituteNaval Postgraduate School833 Dyer Road, Code MOVES, Spanagel 254Monterey, CA 93943-5118(831) [email protected]://www.movesinstitute.org/~augcog/