Title: On the Use of Intelligent Agents as Partners in Training Systems for Complex Tasks*
1On the Use of Intelligent Agentsas Partners in
Training Systemsfor Complex Tasks
- Thomas R. Ioerger, Joe Sims, Richard Volz
- Department of Computer Science
- Texas AM University
- Judson Workman, Wayne Shebilske
- Department of Psychology
- Wright State University
Funds provided by a MURI grant through DoD/AFOSR.
2Complex Tasks, and the Needfor new Training
Methods
- Complex tasks (e.g. operating machinery)
- multiple cognitive components (memory,
perceptual, motor, reasoning/inference...) - novices feel over-whelmed
- limitations of part-task training
- automaticity vs. attention management
- Role for intelligent agents?
- can place agents in simulation environments
- need guiding principles to promote learning
3Previous Work Partner-Based Training
- AIM (Active Interlocked Modeling Shebilske,
1992) - trainees work in pairs (AIM-Dyad)
- each trainee does part of the task together
- importance of context (integration of responses)
- can produce equal training, 100 efficiency gain
- co-presence/social variables not required
- trainees placed in separate rooms
- correlation with intelligence of partner
- Bandura, 1986 modeling
4Automating the Partner with an Intelligent Agent
- Hypothesis Would the training be as effective if
the partner were played by an intelligent agent? - Important pre-requisite a CTA (cognitive task
analysis) - a hierarchical task-decomposition allows
functions to be divided in a natural way
between human and agent partners
5Space Fortress Laboratory Task
- Representative of complex tasks
- has similar perceptual, motor, attention, memory,
and decision-making demands as flying a fighter
jet - continuous control navigation with joystick,
2nd-order thrust control - discrete events firing missles, making bonus
selections with mouse - must learn rules for when to fire, boundaries...
- Large body of previous studies/data
- Multiple Emphasis on Components (MEC) protocol
- transfers to operational setting (attention mgmt)
6P
M
I
MOUSE BUTTONS
JOYSTICK
THE FORTRESS
FORTRESS SHOT
SHIP
BONUS AVAILABLE
A MINE
PNTS CNTRL VLCTY VLNER IFF
INTRVL SPEED SHOTS 200
100 119 0
W 90
70
7Implementation of a Partner Agent
- Implemented decision-making procedures for
automating mouse and joystick - Added if-then-else rules in C source code
- emulate Decision-Making with rules
- Agent simple, but satisfies criteria
- situated, goal-oriented, autonomous
- First version of agent played too perfectly
- Make it play realistically by adding some
delays and imprecision (e.g. in aiming)
8Agent Finite-State Diagrams
Handling the Fortress
Handling Mines
9Experiment 1
- Hypothesis Training with agent improves final
scores - Protocol
- 10 sessions of 10 3-minute trials each (over 4
days) - each session 1/2 hour 8 practice trials, 2 test
trials - Groups
- Control (standard instructionspractice)
- Partner Agent (instructionspractice, alternate
mouse and joystick between trainee and agent) - Participants
- 40 male undegrads at WSU
- lt20 hrs/wk playing video games
10Results of Expt 1
Difference in final scores was significant at
plt0.05 level by paired T-test (with dof38)
t2.33gt2.04
11Breakdown of Scores
12Effect of Level of Simulated Expertise of Agent?
- Results of Expt 1 raises follow-up question What
is the effect of the level of expertise simulated
by the agent? - Can make the agent more or less accurate.
- Recall correlation with partners intelligence
- Is it better to train with an expert? or perhaps
with a partner of matching skill-level?... - novices might have trouble comprehending experts
strategies since struggling to keep up
13Experiment 2
- Hypothesis Different skill-levels of agent
affect trainees performance improvement - Similar design as Expt 1, except 4 Groups
- Control, Novice agent, Intemediate agent, Expert
- Adjust skill-level of agent by fine-tuning
randomness parameters (shot timing, aiming
accur., IFF mistakes) - Gauge to skill levels target groups (empirical)
14Results of Expt 2
Conclusion Training with an expert partner agent
is best.
15Lessons Learned for Future Applications
- Principled approach to using agents in training
systems as partners - cognitive benefits - Requires CTA
- best if high degree of de-coupling
- if greater interaction, agent might have to
cooperate with human by interpreting and
responding to apparent strategies - Desiderata for Agent
- Correctness
- Consistency (necessary for modeling)
- Realism (how to simulate human errors?)
- Exploration (errors lead to unusual situations)