Page 72 - Socially Intelligent Agents Creating Relationships with Computers and Robots

Adapting to Affect and Personality 55

ABAIS ADAPTIVE FRAMEWORK
USER STATE ASSESSMENT IMPACT PREDICTION
TASK
Affect Assessment GENERIC SPECIFIC
- Physiological
- Diagnostic tasks affective state
Affect Affect
- Self report
- KB methods Impact RB Impact RB
Belief Assessment Beliefs
Task / Context - KE techniques individual beliefs Impact RB Beliefs
- KB methods Impact RB
Characteristics
- Diagnostic tasks
specific affect/
belief induced
impact on task
performance
GUI ADAPTATION
STRATEGY SELECTION
DSS GUI task-specific
specific GUI Select Identify compensatory Strategy
adaptation best info additional strategy KB
directives presentation required
strategy information
from DSS Select compensatory
strategies
UI Strategy KB
DECISION SUPPORT
Task Simulation SYSTEM

Figure 6.1. ABAIS Affect-Adaptive Architecture.

3. Adaptive Methodology and Architecture

We developed a methodology designed to compensate for performance bi-
ases caused by users’ affective states and active beliefs [4]. The methodology
consists of four stages: 1) assessing the user’s affective state and performance-
relevant beliefs; 2) identifying their potential impact on performance (e.g., fo-
cus on threatening stimuli); 3) selecting a compensatory strategy (e.g., pre-
sentation of additional information to reduce ambiguity); and 4) implementing
this strategy in terms of specific GUI adaptations (e.g., presenting additional
information, or changing information format to enhance situation awareness).
This methodology was implemented within an architecture: the Affect and
Belief Adaptive Interface System (ABAIS). The ABAIS architecture consists
of four modules, described below, each implementing the corresponding step
of the adaptive methodology (see Figure 6.1): User State Assessment, Impact
Prediction, Strategy Selection, and GUI Adaptation.